Feedback fixes:
1. Revert _convert_vision_content — vision is handled by the vision_analyze
tool, not by converting image blocks inline in conversation messages.
Removed the function and its tests.
2. Add Anthropic to 'hermes model' (cmd_model in main.py):
- Added to provider_labels dict
- Added to providers selection list
- Added _model_flow_anthropic() with Claude Code credential auto-detection,
API key prompting, and model selection from catalog.
3. Wire up Anthropic as a vision-capable auxiliary provider:
- Added _try_anthropic() to auxiliary_client.py using claude-sonnet-4
as the vision model (Claude natively supports multimodal)
- Added to the get_vision_auxiliary_client() auto-detection chain
(after OpenRouter/Nous, before Codex/custom)
Cache tracking note: the Anthropic cache metrics branch in run_agent.py
(cache_read_input_tokens / cache_creation_input_tokens) is in the correct
place — it's response-level parsing, same location as the existing
OpenRouter cache tracking. auxiliary_client.py has no cache tracking.
* fix: use session_key instead of chat_id for adapter interrupt lookups
monitor_for_interrupt() in _run_agent was using source.chat_id to query
the adapter's has_pending_interrupt() and get_pending_message() methods.
But the adapter stores interrupt events under build_session_key(source),
which produces a different string (e.g. 'agent:main:telegram:dm' vs '123456').
This key mismatch meant the interrupt was never detected through the
adapter path, which is the only active interrupt path for all adapter-based
platforms (Telegram, Discord, Slack, etc.). The gateway-level interrupt
path (in dispatch_message) is unreachable because the adapter intercepts
the 2nd message in handle_message() before it reaches dispatch_message().
Result: sending a new message while subagents were running had no effect —
the interrupt was silently lost.
Fix: replace all source.chat_id references in the interrupt-related code
within _run_agent() with the session_key parameter, which matches the
adapter's storage keys.
Also adds regression tests verifying session_key vs chat_id consistency.
* debug: add file-based logging to CLI interrupt path
Temporary instrumentation to diagnose why message-based interrupts
don't seem to work during subagent execution. Logs to
~/.hermes/interrupt_debug.log (immune to redirect_stdout).
Two log points:
1. When Enter handler puts message into _interrupt_queue
2. When chat() reads it and calls agent.interrupt()
This will reveal whether the message reaches the queue and
whether the interrupt is actually fired.
* fix: accept unlisted models with warning instead of rejecting
validate_requested_model() previously hard-rejected any model not found
in the provider's API listing. This was too aggressive — users on higher
plan tiers (e.g. Z.AI Pro/Max) may have access to models not shown in
the public listing (like glm-5 on coding endpoints).
Changes:
- validate_requested_model: accept unlisted models with a warning note
instead of blocking. The model is saved to config and used immediately.
- Z.AI setup: always offer glm-5 in the model list regardless of whether
a coding endpoint was detected. Pro/Max plans support it.
- Z.AI setup detection message: softened from 'GLM-5 is not available'
to 'GLM-5 may still be available depending on your plan tier'
After studying clawdbot (OpenClaw) and OpenCode implementations:
## Beta headers
- Add interleaved-thinking-2025-05-14 and fine-grained-tool-streaming-2025-05-14
as common betas (sent with ALL auth types, not just OAuth)
- OAuth tokens additionally get oauth-2025-04-20
- API keys now also get the common betas (previously got none)
## Vision/image support
- Add _convert_vision_content() to convert OpenAI multimodal format
(image_url blocks) to Anthropic format (image blocks with base64/url source)
- Handles both data: URIs (base64) and regular URLs
## Role alternation enforcement
- Anthropic strictly rejects consecutive same-role messages (400 error)
- Add post-processing step that merges consecutive user/assistant messages
- Handles string, list, and mixed content types during merge
## Tool choice support
- Add tool_choice parameter to build_anthropic_kwargs()
- Maps OpenAI values: auto→auto, required→any, none→omit, name→tool
## Cache metrics tracking
- Anthropic uses cache_read_input_tokens / cache_creation_input_tokens
(different from OpenRouter's prompt_tokens_details.cached_tokens)
- Add api_mode-aware branch in run_agent.py cache stats logging
## Credential refresh on 401
- On 401 error during anthropic_messages mode, re-read credentials
via resolve_anthropic_token() (picks up refreshed Claude Code tokens)
- Rebuild client if new token differs from current one
- Follows same pattern as Codex/Nous 401 refresh handlers
## Tests
- 44 adapter tests (8 new: vision conversion, role alternation, tool choice)
- Updated beta header tests to verify new structure
- Full suite: 3198 passed, 0 regressions
* fix: use session_key instead of chat_id for adapter interrupt lookups
monitor_for_interrupt() in _run_agent was using source.chat_id to query
the adapter's has_pending_interrupt() and get_pending_message() methods.
But the adapter stores interrupt events under build_session_key(source),
which produces a different string (e.g. 'agent:main:telegram:dm' vs '123456').
This key mismatch meant the interrupt was never detected through the
adapter path, which is the only active interrupt path for all adapter-based
platforms (Telegram, Discord, Slack, etc.). The gateway-level interrupt
path (in dispatch_message) is unreachable because the adapter intercepts
the 2nd message in handle_message() before it reaches dispatch_message().
Result: sending a new message while subagents were running had no effect —
the interrupt was silently lost.
Fix: replace all source.chat_id references in the interrupt-related code
within _run_agent() with the session_key parameter, which matches the
adapter's storage keys.
Also adds regression tests verifying session_key vs chat_id consistency.
* debug: add file-based logging to CLI interrupt path
Temporary instrumentation to diagnose why message-based interrupts
don't seem to work during subagent execution. Logs to
~/.hermes/interrupt_debug.log (immune to redirect_stdout).
Two log points:
1. When Enter handler puts message into _interrupt_queue
2. When chat() reads it and calls agent.interrupt()
This will reveal whether the message reaches the queue and
whether the interrupt is actually fired.
Follow-up to PR #862 (local skills classification by arceus77-7):
- Remove unnecessary isinstance guard on _read_manifest() return value —
it always returns Dict[str, str], so set() on it suffices.
- Extract repeated hub-dir monkeypatching into a shared pytest fixture (hub_env).
- Add three_source_env fixture for source-classification tests.
- Add _read_manifest monkeypatch to test_do_list_initializes_hub_dir
(was fragile — relied on empty skills list masking the real manifest).
- Add test coverage for --source hub and --source builtin filters.
- Extract _capture() helper to reduce console/StringIO boilerplate.
5 tests, all green.
When a dangerous command is detected and the user is prompted for
approval, long commands are truncated (80 chars in fallback, 70 chars
in the TUI). Users had no way to see the full command before deciding.
This adds a 'View full command' option across all approval interfaces:
- CLI fallback (tools/approval.py): [v]iew option in the prompt menu.
Shows the full command and re-prompts for approval decision.
- CLI TUI (cli.py): 'Show full command' choice in the arrow-key
selection panel. Expands the command display in-place and removes
the view option after use.
- CLI callbacks (callbacks.py): 'view' choice added to the list when
the command exceeds 70 characters.
- Gateway (gateway/run.py): 'full', 'show', 'view' responses reveal
the complete command while keeping the approval pending.
Includes 7 new tests covering view-then-approve, view-then-deny,
short command fallthrough, and double-view behavior.
Closes community feedback about the 80-char cap on dangerous commands.
* fix: /reasoning command output ordering, display, and inline think extraction
Three issues with the /reasoning command:
1. Output interleaving: The command echo used print() while feedback
used _cprint(), causing them to render out-of-order under
prompt_toolkit's patch_stdout. Changed echo to use _cprint() so
all output renders through the same path in correct order.
2. Reasoning display not working: /reasoning show toggled a flag
but reasoning never appeared for models that embed thinking in
inline <think> blocks rather than structured API fields. Added
fallback extraction in _build_assistant_message to capture
<think> block content as reasoning when no structured reasoning
fields (reasoning, reasoning_content, reasoning_details) are
present. This feeds into both the reasoning callback (during
tool loops) and the post-response reasoning box display.
3. Feedback clarity: Added checkmarks to confirm actions, persisted
show/hide to config (was session-only before), and aligned the
status display for readability.
Tests: 7 new tests for inline think block extraction (41 total).
* feat: add /reasoning command to gateway (Telegram/Discord/etc)
The /reasoning command only existed in the CLI — messaging platforms
had no way to view or change reasoning settings. This adds:
1. /reasoning command handler in the gateway:
- No args: shows current effort level and display state
- /reasoning <level>: sets reasoning effort (none/low/medium/high/xhigh)
- /reasoning show|hide: toggles reasoning display in responses
- All changes saved to config.yaml immediately
2. Reasoning display in gateway responses:
- When show_reasoning is enabled, prepends a 'Reasoning' block
with the model's last_reasoning content before the response
- Collapses long reasoning (>15 lines) to keep messages readable
- Uses last_reasoning from run_conversation result dict
3. Plumbing:
- Added _show_reasoning attribute loaded from config at startup
- Propagated last_reasoning through _run_agent return dict
- Added /reasoning to help text and known_commands set
- Uses getattr for _show_reasoning to handle test stubs
4 test files spawn real processes or make live API calls that hang
indefinitely in batch/CI runs. Skip them with pytestmark:
- tests/tools/test_code_execution.py (subprocess spawns)
- tests/tools/test_file_tools_live.py (live LocalEnvironment)
- tests/test_413_compression.py (blocks on process)
- tests/test_agent_loop_tool_calling.py (live OpenRouter API calls)
Also added global 30s signal.alarm timeout in conftest.py as a safety
net, and removed stale nous-api test that hung on OAuth browser login.
Suite now runs in ~55s with no hangs.
- Remove test_nous_api_setup_preserves_model_provider_metadata (nous-api
provider no longer exists, test selected Nous OAuth which hangs waiting
for browser login)
- Fix test_nous_oauth_setup prompt_choice index: 1→0 (Nous Portal is
now first option after nous-api removal)
- gateway/run.py: Take main's _resolve_gateway_model() helper
- hermes_cli/setup.py: Re-apply nous-api removal after merge brought
it back. Fix provider_idx offset (Custom is now index 3, not 4).
- tests/hermes_cli/test_setup.py: Fix custom setup test index (3→4)
Both /model and /provider now show the same unified display:
Current: anthropic/claude-opus-4.6 via OpenRouter
Authenticated providers & models:
[openrouter] ← active
anthropic/claude-opus-4.6 ← current
anthropic/claude-sonnet-4.5
...
[nous]
claude-opus-4-6
gemini-3-flash
...
[openai-codex]
gpt-5.2-codex
gpt-5.1-codex-mini
...
Not configured: Z.AI / GLM, Kimi / Moonshot, ...
Switch model: /model <model-name>
Switch provider: /model <provider>:<model-name>
Example: /model nous:claude-opus-4-6
Users can see all authenticated providers and their models at a glance,
making it easy to switch mid-conversation.
Also added curated model lists for Nous Portal and OpenAI Codex to
hermes_cli/models.py.
Model selection now comes exclusively from config.yaml (set via
'hermes model' or 'hermes setup'). The LLM_MODEL env var is no longer
read or written anywhere in production code.
Why: env vars are per-process/per-user and would conflict in
multi-agent or multi-tenant setups. Config.yaml is file-based and
can be scoped per-user or eventually per-session.
Changes:
- cli.py: Read model from CLI_CONFIG only, not LLM_MODEL/OPENAI_MODEL
- hermes_cli/auth.py: _save_model_choice() no longer writes LLM_MODEL
to .env
- hermes_cli/setup.py: Remove 12 save_env_value('LLM_MODEL', ...)
calls from all provider setup flows
- gateway/run.py: Remove LLM_MODEL fallback (HERMES_MODEL still works
for gateway process runtime)
- cron/scheduler.py: Same
- agent/auxiliary_client.py: Remove LLM_MODEL from custom endpoint
model detection
Phase 2 of the provider router migration — route the main agent's
client construction and fallback activation through
resolve_provider_client() instead of duplicated ad-hoc logic.
run_agent.py:
- __init__: When no explicit api_key/base_url, use
resolve_provider_client(provider, raw_codex=True) for client
construction. Explicit creds (from CLI/gateway runtime provider)
still construct directly.
- _try_activate_fallback: Replace _resolve_fallback_credentials and
its duplicated _FALLBACK_API_KEY_PROVIDERS / _FALLBACK_OAUTH_PROVIDERS
dicts with a single resolve_provider_client() call. The router
handles all provider types (API-key, OAuth, Codex) centrally.
- Remove _resolve_fallback_credentials method and both fallback dicts.
agent/auxiliary_client.py:
- Add raw_codex parameter to resolve_provider_client(). When True,
returns the raw OpenAI client for Codex providers instead of wrapping
in CodexAuxiliaryClient. The main agent needs this for direct
responses.stream() access.
3251 passed, 2 pre-existing unrelated failures.
Update 14 test files to use the new call_llm/async_call_llm mock
patterns instead of the old get_text_auxiliary_client/
get_vision_auxiliary_client tuple returns.
- vision_tools tests: mock async_call_llm instead of _aux_async_client
- browser tests: mock call_llm instead of _aux_vision_client
- flush_memories tests: mock call_llm instead of get_text_auxiliary_client
- session_search tests: mock async_call_llm with RuntimeError
- mcp_tool tests: fix whitelist model config, use side_effect for
multi-response tests
- auxiliary_config_bridge: update for model=None (resolved in router)
3251 passed, 2 pre-existing unrelated failures.
Add centralized call_llm() and async_call_llm() functions that own the
full LLM request lifecycle:
1. Resolve provider + model from task config or explicit args
2. Get or create a cached client for that provider
3. Format request args (max_tokens handling, provider extra_body)
4. Make the API call with max_tokens/max_completion_tokens retry
5. Return the response
Config: expanded auxiliary section with provider:model slots for all
tasks (compression, vision, web_extract, session_search, skills_hub,
mcp, flush_memories). Config version bumped to 7.
Migrated all auxiliary consumers:
- context_compressor.py: uses call_llm(task='compression')
- vision_tools.py: uses async_call_llm(task='vision')
- web_tools.py: uses async_call_llm(task='web_extract')
- session_search_tool.py: uses async_call_llm(task='session_search')
- browser_tool.py: uses call_llm(task='vision'/'web_extract')
- mcp_tool.py: uses call_llm(task='mcp')
- skills_guard.py: uses call_llm(provider='openrouter')
- run_agent.py flush_memories: uses call_llm(task='flush_memories')
Tests updated for context_compressor and MCP tool. Some test mocks
still need updating (15 remaining failures from mock pattern changes,
2 pre-existing).
Prevent stale Honcho tool exposure in context/local modes, restore reliable async write retry behavior, and ensure SOUL.md migration uploads target the AI peer instead of the user peer. Also align Honcho CLI key checks with host-scoped apiKey resolution and lock the fixes with regression tests.
Made-with: Cursor
When a new user runs 'hermes setup' for the first time and ~/.openclaw/
exists, the wizard now asks if they want to import their OpenClaw data
before API/tool configuration begins.
If accepted, the existing migration script from optional-skills/ is
loaded dynamically and run with the 'full' preset — importing settings,
memories, skills, API keys, and platform configs. Config is reloaded
afterward so imported values (like API keys) are available for the
remaining setup steps.
The migration is only offered on first-time setup (not returning users)
and handles errors gracefully without blocking setup completion.
Closes#829
Matches Hermes' native session naming (title if set, otherwise
session-scoped). Not a breaking change -- no memory data is lost,
old sessions remain in Honcho.
When hermes-agent runs as a systemd service, Docker container, or
headless daemon, the stdout pipe can become unavailable (idle timeout,
buffer exhaustion, socket reset). Any print() call then raises
OSError: [Errno 5] Input/output error, crashing run_conversation()
and causing cron jobs to fail.
Rather than wrapping individual print() calls (68 in run_conversation
alone), this adds a transparent _SafeWriter wrapper installed once at
the start of run_conversation(). It delegates all writes to the real
stdout and silently catches OSError. Zero overhead on the happy path,
comprehensive coverage of all print calls including future ones.
Fixes#845
Co-authored-by: J0hnLawMississippi <J0hnLawMississippi@users.noreply.github.com>
- Forum parent channel IDs now match free-response list (add a forum
channel ID and all its threads respond without mention)
- Better thread chat names: 'Guild / forum / thread' for forum threads
- Add discord.require_mention and discord.free_response_channels to
config.yaml (bridged to env vars, env vars still override)
- Keep require_mention defaulting to true (safe for shared servers)
Cherry-picked from PR #867 by insecurejezza with default fix and
config.yaml integration.
Co-authored-by: insecurejezza <insecurejezza@users.noreply.github.com>
Address merge-blocking review feedback by removing unsafe signal handler overrides, wiring next-turn Honcho prefetch, restoring per-directory session defaults, and exposing all Honcho tools to the model surface. Also harden prefetch cache access with public thread-safe accessors and remove duplicate browser cleanup code.
Made-with: Cursor
Three new tests for the naive timestamp fix (PR #807):
- test_ensure_aware_naive_preserves_absolute_time: verifies UTC equivalent
is preserved when interpreting naive datetimes as system-local time
- test_ensure_aware_normalizes_aware_to_hermes_tz: verifies already-aware
datetimes are normalized to Hermes tz without shifting the instant
- test_ensure_aware_due_job_not_skipped_when_system_ahead: end-to-end
regression test for the original bug scenario
_discover_one() caught all exceptions and returned [], making
asyncio.gather(return_exceptions=True) redundant. The
isinstance(result, Exception) branch in _discover_all() was dead
code, so failed_count was always 0. This caused:
- No summary printed when all servers fail (silent failure)
- ok_servers always equaling total_servers (misleading count)
- Unused variables transport_desc and transport_type
Fix: let exceptions propagate to gather() so failed_count increments
correctly. Move per-server failure logging to _discover_all(). Remove
dead variables.
Two fixes for context overflow handling:
1. Proactive compression after tool execution: The compression check now
estimates the next prompt size using real token counts from the last API
response (prompt_tokens + completion_tokens) plus a conservative estimate
of newly appended tool results (chars // 3 for JSON-heavy content).
Previously, should_compress() only checked last_prompt_tokens which
didn't account for tool results — so a 130k prompt + 100k chars of tool
output would pass the 140k threshold check but fail the 200k API limit.
2. Safety net: Added 'prompt is too long' to context-length error detection
phrases. Anthropic returns 'prompt is too long: N tokens > M maximum'
on HTTP 400, which wasn't matched by existing phrases. This ensures
compression fires even if the proactive check underestimates.
Fixes#813
- test_agent_loop_tool_calling.py: import atroposlib at module level
to trigger skip (environments.agent_loop is now importable without
atroposlib due to __init__.py graceful fallback)
- test_modal_sandbox_fixes.py: skip TestToolResolution tests when
minisweagent not installed
- environments/__init__.py: try/except on atroposlib imports so
submodules like tool_call_parsers remain importable standalone
- test_agent_loop.py, test_tool_call_parsers.py,
test_managed_server_tool_support.py: skip at module level when
atroposlib is missing
Guard all test files that import from environments/ or atroposlib
with try/except + pytest.skip(allow_module_level=True) so they
gracefully skip instead of crashing when deps aren't available.
Tests hit a real vLLM server (Qwen/Qwen3-4B-Thinking-2507) via
ManagedServer Phase 2. Auto-skip if server isn't running.
Tests verify:
- Single tool call through full agent loop
- Multi-tool calls across turns
- ManagedServer produces SequenceNodes with tokens/logprobs
- Direct response without tools
- Thinking model produces <think> blocks
Also adds fallback parser in agent_loop.py: when ManagedServer's
ToolCallTranslator can't parse (vLLM not installed), hermes-agent's
standalone parsers extract <tool_call> tags from raw content.
Real LLM calls via OpenRouter using stepfun/step-3.5-flash:free (zero cost).
Falls back to paid models if free model is unavailable.
Tests: single tool call, multi-tool single turn, multi-turn chains,
unknown tool rejection, max_turns limit, direct response (no tools),
tool error handling, AgentResult structure, conversation history.
The weaker assertion (r.exc_info is not None) passes even when
exc_info is (None, None, None). Check r.exc_info[0] is not None
to verify actual exception info is present.
The _aux_async_client mock was already applied on main.
Co-authored-by: OutThisLife <nickolasgustafsson@gmail.com>
Allow users to interact with Hermes by sending and receiving emails.
Uses IMAP polling for incoming messages and SMTP for replies with
proper threading (In-Reply-To, References headers).
Integrates with all 14 gateway extension points: config, adapter
factory, authorization, send_message tool, cron delivery, toolsets,
prompt hints, channel directory, setup wizard, status display, and
env example.
65 tests covering config, parsing, dispatch, threading, IMAP fetch,
SMTP send, attachments, and all integration points.
Adds delegation.model and delegation.provider config fields so subagents
can run on a completely different provider:model pair than the parent agent.
When delegation.provider is set, the system resolves the full credential
bundle (base_url, api_key, api_mode) via resolve_runtime_provider() —
the same path used by CLI/gateway startup. This means all configured
providers work out of the box: openrouter, nous, zai, kimi-coding,
minimax, minimax-cn.
Key design decisions:
- Provider resolution uses hermes_cli.runtime_provider (single source of
truth for credential resolution across CLI, gateway, cron, and now
delegation)
- When only delegation.model is set (no provider), the model name changes
but parent credentials are inherited (for switching models within the
same provider like OpenRouter)
- When delegation.provider is set, full credentials are resolved
independently — enabling cross-provider delegation (e.g. parent on
Nous Portal, subagents on OpenRouter)
- Clear error messages if provider resolution fails (missing API key,
unknown provider name)
- _load_config() now falls back to hermes_cli.config.load_config() for
gateway/cron contexts where CLI_CONFIG is unavailable
Based on PR #791 by 0xbyt4 (closes#609), reworked to use proper
provider credential resolution instead of passing provider as metadata.
Co-authored-by: 0xbyt4 <0xbyt4@users.noreply.github.com>
Combined implementation of reasoning management:
- /reasoning Show current effort level and display state
- /reasoning <level> Set reasoning effort (none, low, medium, high, xhigh)
- /reasoning show|on Show model thinking/reasoning in output
- /reasoning hide|off Hide model thinking/reasoning from output
Effort level changes persist to config and force agent re-init.
Display toggle updates the agent callback dynamically without re-init.
When display is enabled:
- Intermediate reasoning shown as dim [thinking] lines during tool loops
- Final reasoning shown in a bordered box above the response
- Long reasoning collapsed (5 lines intermediate, 10 lines final)
Also adds:
- reasoning_callback parameter to AIAgent
- last_reasoning in run_conversation result dict
- show_reasoning config option (display section, default: false)
- Display section in /config output
- 34 tests covering both features
Combines functionality from PR #789 and PR #790.
Co-authored-by: Aum Desai <Aum08Desai@users.noreply.github.com>
Co-authored-by: 0xbyt4 <35742124+0xbyt4@users.noreply.github.com>
Adds defensive guard against empty/None/missing choices in SamplingHandler.__call__
before accessing response.choices[0]. Returns proper ErrorData instead of crashing
with IndexError/TypeError on content filtering, provider errors, or rate limits.
Authored by 0xbyt4.
Co-authored-by: 0xbyt4 <0xbyt4@users.noreply.github.com>
Adds tool_choice, parallel_tool_calls, and prompt_cache_key to the
Codex Responses API request kwargs — matching what the official Codex
CLI sends.
- tool_choice: 'auto' — enables the model to proactively call tools.
Without this, the model may default to not using tools, which explains
reports of the agent claiming it lacks shell access (#747).
- parallel_tool_calls: True — allows the model to issue multiple tool
calls in a single turn for efficiency.
- prompt_cache_key: session_id — enables server-side prompt caching
across turns in the same session, reducing latency and cost.
Refs #747
Isolate Telegram forum topic sessions — each topic gets its own independent session key, history, and interrupt tracking. Progress, hygiene, and cron messages all route to the correct topic.
Four cleanups to code merged today:
1. New hermes_cli/curses_ui.py — shared curses_checklist() used by both
hermes tools and hermes skills. Eliminates ~140 lines of near-identical
curses code (scrolling, key handling, color setup, numbered fallback).
2. Fix _find_all_skills() perf — was calling load_config() per skill
(~100+ YAML parses). Now loads disabled set once via
_get_disabled_skill_names() and does a set lookup.
3. Eliminate _list_all_skills_unfiltered() duplication — _find_all_skills()
now accepts skip_disabled=True for the config UI, removing 30 lines
of copy-pasted discovery logic from skills_config.py.
4. Fix fragile label round-trip in skills_command — was building label
strings, passing to checklist, then mapping labels back to skill names
(collision-prone). Now works with indices directly, like tools_config.
Authored by teyrebaz33. Closes#643.
- /personality none/default/neutral clears system prompt overlay
- Dict format personalities with description, tone, style fields
- Works in both CLI and gateway
- 18 tests
Add /background <prompt> to the gateway, allowing users on Telegram,
Discord, Slack, etc. to fire off a prompt in a separate agent session.
The result is delivered back to the same chat when done, without
modifying the active conversation history.
Implementation:
- _handle_background_command: validates input, spawns asyncio task
- _run_background_task: creates AIAgent in executor thread, delivers
result (text, images, media files) back via the platform adapter
- Inherits model, toolsets, provider routing from gateway config
- Error handling with user-visible failure messages
Also adds /background to hermes_cli/commands.py registry so it
appears in /help and autocomplete.
Tests: 15 new tests covering usage, task creation, uniqueness,
multi-platform, error paths, and help/autocomplete integration.
Fixes#898 — Python 3.11 changed argparse to raise an exception on
duplicate subparser names (CPython #94331). The 'skills' name was
registered twice: once for Skills Hub and once for skills config.
Changes:
- Remove duplicate 'skills' subparser registration
- Add 'config' as a sub-action under the existing 'hermes skills' command
- Route 'hermes skills config' to skills_config module
- Add regression test to catch future duplicates
Migration: 'hermes skills' (config) is now 'hermes skills config'
Two-tier warning system that nudges the LLM as it approaches
max_iterations, injected into the last tool result JSON rather
than as a separate system message:
- Caution (70%): {"_budget_warning": "[BUDGET: 42/60...]"}
- Warning (90%): {"_budget_warning": "[BUDGET WARNING: 54/60...]"}
For JSON tool results, adds a _budget_warning field to the existing
dict. For plain text results, appends the warning as text.
Key properties:
- No system messages injected mid-conversation
- No changes to message structure
- Prompt cache stays valid
- Configurable thresholds (0.7 / 0.9)
- Can be disabled: _budget_pressure_enabled = False
Inspired by PR #421 (@Bartok9) and issue #414.
8 tests covering thresholds, edge cases, JSON and text injection.
Replaces head-only stdout capture with a two-buffer approach (40% head,
60% tail rolling window) so scripts that print() their final results
at the end never lose them. Adds truncation notice between sections.
Cherry-picked from PR #755, conflict resolved (test file additions).
3 new tests for short output, head+tail preservation, and notice format.
Adds configurable bot message filtering via DISCORD_ALLOW_BOTS env var:
- 'none' (default): ignore all bot messages
- 'mentions': accept bots only when they @mention us
- 'all': accept all bot messages
Includes 8 tests.
Authored by teyrebaz33. Adds config-driven quick commands that execute
shell commands without invoking the LLM — zero token usage, works from
Telegram/Discord/Slack/etc. Closes#744.
Follow-up to 58dbd81 — ensures smooth transition for existing users:
- Backward compat: old session files without last_prompt_tokens
default to 0 via data.get('last_prompt_tokens', 0)
- /compress, /undo, /retry: reset last_prompt_tokens to 0 after
rewriting transcripts (stale token counts would under-report)
- Auto-compression hygiene: reset last_prompt_tokens after rewriting
- update_session: use None sentinel (not 0) as default so callers
can explicitly reset to 0 while normal calls don't clobber
- 6 new tests covering: default value, serialization roundtrip,
old-format migration, set/reset/no-change semantics
- /reset: new SessionEntry naturally gets last_prompt_tokens=0
2942 tests pass.
Adds two tests for _handle_retry_command: verifies /retry returns the
agent response (not None), and verifies graceful handling when no
previous message exists.
Cherry-picked from PR #731 by teyrebaz33. Regression coverage for
the fix merged in PR #441.
Co-authored-by: teyrebaz33 <teyrebaz33@users.noreply.github.com>
Authored by dmahan93. Adds HERMES_YOLO_MODE env var and --yolo CLI flag
to auto-approve all dangerous command prompts.
Post-merge: renamed --fuck-it-ship-it to --yolo for brevity,
resolved conflict with --checkpoints flag.
On macOS, Docker Desktop installs the CLI to /usr/local/bin/docker, but
when Hermes runs as a gateway service (launchd) or in other non-login
contexts, /usr/local/bin is often not in PATH. This causes the Docker
requirements check to fail with 'No such file or directory: docker' even
though docker works fine from the user's terminal.
Add find_docker() helper that uses shutil.which() first, then probes
common Docker Desktop install paths on macOS (/usr/local/bin,
/opt/homebrew/bin, Docker.app bundle). The resolved path is cached and
passed to mini-swe-agent via its 'executable' parameter.
- tools/environments/docker.py: add find_docker(), use it in
_storage_opt_supported() and pass to _Docker(executable=...)
- tools/terminal_tool.py: use find_docker() in requirements check
- tests/tools/test_docker_find.py: 4 tests (PATH, fallback, not found, cache)
2877 tests pass.
Shows an immediate status message and braille spinner for slow slash
commands (/skills search|browse|inspect|install, /reload-mcp). Makes
input read-only while the command runs so the CLI doesn't appear frozen.
Cherry-picked from PR #714 by vilkasdev, rebased onto current main
with conflict resolution and bug fix (get_hint_text duplicate return).
Fixes#636
Co-authored-by: vilkasdev <vilkasdev@users.noreply.github.com>
Two related bugs prevented users from reliably switching providers:
1. OPENAI_BASE_URL poisoning OpenRouter resolution: When a user with a
custom endpoint ran /model openrouter:model, _resolve_openrouter_runtime
picked up OPENAI_BASE_URL instead of the OpenRouter URL, causing model
validation to probe the wrong API and reject valid models.
Fix: skip OPENAI_BASE_URL when requested_provider is explicitly
'openrouter'.
2. Provider never saved to config: _save_model_choice() could save
config.model as a plain string. All five _model_flow_* functions then
checked isinstance(model, dict) before writing the provider — which
silently failed on strings. With no provider in config, auto-detection
would pick up stale credentials (e.g. Codex desktop app) instead of
the user's explicit choice.
Fix: _save_model_choice() now always saves as dict format. All flow
functions also normalize string->dict as a safety net before writing
provider.
Adds 4 regression tests. 2873 tests pass.
Follow-up to PR #716 (0xbyt4):
- Log the third remaining silent except-pass in scheduler (prefill
messages JSON parse failure)
- Fix test mock: run → run_conversation (matches actual agent API)
- Remove unused imports (asyncio, AsyncMock)
- Add test for prefill_messages parse failure logging
The 4 early-return paths in _spawn_training_run (API exit, trainer
exit, env not found, env exit) were doing manual process.terminate()
or returning without cleanup, leaking open log file handles. Now all
paths call _stop_training_run() which handles both process termination
and file handle closure.
Also adds 12 tests for _stop_training_run covering file handle
cleanup, process termination, status transitions, and edge cases.
Inspired by PR #715 (0xbyt4) which identified the early-return issue.
Core file handle fix was already on main via e28dc13 (memosr.eth).
Follow-up to PR #705 (merged from 0xbyt4). Addresses several issues:
1. CONSECUTIVE-ONLY TRACKING: Redesigned the read/search tracker to only
warn/block on truly consecutive identical calls. Any other tool call
in between (write, patch, terminal, etc.) resets the counter via
notify_other_tool_call(), called from handle_function_call() in
model_tools.py. This prevents false blocks in read→edit→verify flows.
2. THRESHOLD ADJUSTMENT: Warn on 3rd consecutive (was 2nd), block on
4th+ consecutive (was 3rd+). Gives the model more room before
intervening.
3. TUPLE UNPACKING BUG: Fixed get_read_files_summary() which crashed on
search keys (5-tuple) when trying to unpack as 3-tuple. Now uses a
separate read_history set that only tracks file reads.
4. WEB_EXTRACT DOCSTRING: Reverted incorrect removal of 'title' from
web_extract return docs in code_execution_tool.py — the field IS
returned by web_tools.py.
5. TESTS: Rewrote test_read_loop_detection.py (35 tests) to cover
consecutive-only behavior, notify_other_tool_call, interleaved
read/search, and summary-unaffected-by-searches.
Three separate code paths all wrote to the same SQLite state.db with
no deduplication, inflating session transcripts by 3-4x:
1. _log_msg_to_db() — wrote each message individually after append
2. _flush_messages_to_session_db() — re-wrote ALL new messages at
every _persist_session() call (~18 exit points), with no tracking
of what was already written
3. gateway append_to_transcript() — wrote everything a third time
after the agent returned
Since load_transcript() prefers SQLite over JSONL, the inflated data
was loaded on every session resume, causing proportional token waste.
Fix:
- Remove _log_msg_to_db() and all 16 call sites (redundant with flush)
- Add _last_flushed_db_idx tracking in _flush_messages_to_session_db()
so repeated _persist_session() calls only write truly new messages
- Reset flush cursor on compression (new session ID)
- Add skip_db parameter to SessionStore.append_to_transcript() so the
gateway skips SQLite writes when the agent already persisted them
- Gateway now passes skip_db=True for agent-managed messages, still
writes to JSONL as backup
Verified: a 12-message CLI session with tool calls produces exactly
12 SQLite rows with zero duplicates (previously would be 36-48).
Tests: 9 new tests covering flush deduplication, skip_db behavior,
compression reset, and initialization. Full suite passes (2869 tests).
- Add agent/embeddings.py with Embedder protocol, FastEmbedEmbedder, OpenAIEmbedder
- Factory function get_embedder() reads provider from config.yaml embeddings section
- Lazy initialization — no startup impact, model loaded on first embed call
- cosine_similarity() and cosine_similarity_matrix() utility functions included
- Add fastembed as optional dependency in pyproject.toml
- 30 unit tests, all passing
Closes#675
Completes the fix started in 8318a51 — handle_function_call() accepted
enabled_tools but run_agent.py never passed it. Now both call sites in
_execute_tool_calls() pass self.valid_tool_names, so each agent session
uses its own tool list instead of the process-global
_last_resolved_tool_names (which subagents can overwrite).
Also simplifies the redundant ternary in code_execution_tool.py:
sandbox_tools is already computed correctly (intersection with session
tools, or full SANDBOX_ALLOWED_TOOLS as fallback), so the conditional
was dead logic.
Inspired by PR #663 (JasonOA888). Closes#662.
Tests: 2857 passed.
Add support for using Nous Portal via a direct API key, mirroring
how OpenRouter and other API-key providers work. This gives users a
simpler alternative to the OAuth device-code flow when they already
have a Nous API key.
Changes:
- Add 'nous-api' to PROVIDER_REGISTRY as an api_key provider
pointing to https://inference-api.nousresearch.com/v1
- Add NOUS_API_KEY and NOUS_BASE_URL to OPTIONAL_ENV_VARS
- Add NOUS_API_BASE_URL / NOUS_API_CHAT_URL to hermes_constants
- Add 'Nous Portal API key' as first option in setup wizard
- Add provider aliases (nous_api, nousapi, nous-portal-api)
- Add test for nous-api runtime provider resolution
Closes#644
Move all new tests (schema, env filtering, edge cases, interrupt) into
the existing test_code_execution.py instead of a separate file.
Delete the now-redundant test_code_execution_schema.py.
build_execute_code_schema(set()) produced "from hermes_tools import , ..."
in the code property description — invalid Python syntax shown to the model.
This triggers when a user enables only the code_execution toolset without
any of the sandbox-allowed tools (e.g. `hermes tools code_execution`),
because SANDBOX_ALLOWED_TOOLS & {"execute_code"} = empty set.
Also adds 29 unit tests covering build_execute_code_schema, environment
variable filtering, execute_code edge cases, and interrupt handling.
Authored by tripledoublev.
After context compression on 413/400 errors, the inner retry loop was
reusing the stale pre-compression api_messages payload. Fix breaks out
of the inner retry loop so the outer loop rebuilds api_messages from
the now-compressed messages list. Adds regression test verifying the
second request actually contains the compressed payload.
Add display.background_process_notifications config option to control
how chatty the gateway process watcher is when using
terminal(background=true, check_interval=...) from messaging platforms.
Modes:
- all: running-output updates + final message (default, current behavior)
- result: only the final completion message
- error: only the final message when exit code != 0
- off: no watcher messages at all
Also supports HERMES_BACKGROUND_NOTIFICATIONS env var override.
Includes 12 tests (5 config loading + 7 watcher behavior).
Inspired by @PeterFile's PR #593. Closes#592.
Automatic filesystem snapshots before destructive file operations,
with user-facing rollback. Inspired by PR #559 (by @alireza78a).
Architecture:
- Shadow git repos at ~/.hermes/checkpoints/{hash}/ via GIT_DIR
- CheckpointManager: take/list/restore, turn-scoped dedup, pruning
- Transparent — the LLM never sees it, no tool schema, no tokens
- Once per turn — only first write_file/patch triggers a snapshot
Integration:
- Config: checkpoints.enabled + checkpoints.max_snapshots
- CLI flag: hermes --checkpoints
- Trigger: run_agent.py _execute_tool_calls() before write_file/patch
- /rollback slash command in CLI + gateway (list, restore by number)
- Pre-rollback snapshot auto-created on restore (undo the undo)
Safety:
- Never blocks file operations — all errors silently logged
- Skips root dir, home dir, dirs >50K files
- Disables gracefully when git not installed
- Shadow repo completely isolated from project git
Tests: 35 new tests, all passing (2798 total suite)
Docs: feature page, config reference, CLI commands reference
Authored by unmodeled-tyler. Adds openclaw-migration skill to optional-skills/
with migration script, SKILL.md, and 7 tests. Also improves clarify/approval
panel rendering with dynamic width calculation.
Authored by 0xbyt4. Fixes #N/A (no linked issue).
- Sanitize user input before FTS5 MATCH to prevent OperationalError on
special characters (C++, unbalanced quotes, dangling operators, etc.)
- Close SessionDB connection in mirror._append_to_sqlite() via finally block
- Added tests for both fixes
Platform.SIGNAL was missing from default_toolset_map and platform_config_key
in gateway/run.py, causing Signal to silently fall back to hermes-telegram
toolset (same bug as HomeAssistant, fixed in PR #538).
Also updates:
- tests/test_toolsets.py: include hermes-signal and hermes-homeassistant in
the platform core-tools consistency check
- cli-config.yaml.example: document signal and homeassistant platform keys
When a user runs `hermes -w -c Pokemon Agent Dev` without quoting the
session name, argparse would fail with:
error: argument command: invalid choice: 'Agent'
This is because argparse parses `-c Pokemon` (consuming one token via
nargs='?'), then sees 'Agent' and tries to match it as a subcommand.
Fix: add _coalesce_session_name_args() that pre-processes sys.argv before
argparse, joining consecutive non-flag, non-subcommand tokens after -c or
-r into a single argument. This makes both quoted and unquoted multi-word
session names work transparently.
Includes 17 tests covering all edge cases: multi-word names, single-word,
bare flags, flag ordering, subcommand boundaries, and passthrough.
Complements PR #453 by 0xbyt4. Adds isinstance(dict) guard in
run_agent.py to catch cases where json.loads returns non-dict
(e.g. null, list, string) before they reach downstream code.
Also adds 15 tests for build_tool_preview covering None args,
empty dicts, known/unknown tools, fallback keys, truncation,
and all special-cased tools (process, todo, memory, session_search).
Validate that the username portion of ~username paths contains only
valid characters (alphanumeric, dot, hyphen, underscore) before passing
to shell echo for expansion. Previously, paths like '~; rm -rf /'
would be passed unquoted to self._exec(f'echo {path}'), allowing
arbitrary command execution.
The approach validates the username rather than using shlex.quote(),
which would prevent tilde expansion from working at all since
echo '~user' outputs the literal string instead of expanding it.
Added tests for injection blocking and valid ~username/path expansion.
Credit to @alireza78a for reporting (PR #442, issue #442).
SamplingHandler.__call__ accessed response.choices[0] without checking
if the list was non-empty. LLM APIs can return empty choices on content
filtering, provider errors, or rate limits, causing an unhandled
IndexError that propagates to the MCP SDK and may crash the connection.
Add a defensive guard that returns a proper ErrorData when choices is
empty, None, or missing. Includes three test cases covering all
variants.
Vision auto-mode previously only tried OpenRouter, Nous, and Codex
for multimodal — deliberately skipping custom endpoints with the
assumption they 'may not handle vision input.' This caused silent
failures for users running local multimodal models (Qwen-VL, LLaVA,
Pixtral, etc.) without any cloud API keys.
Now custom endpoints are tried as a last resort in auto mode. If the
model doesn't support vision, the API call fails gracefully — but
users with local vision models no longer need to manually set
auxiliary.vision.provider: main in config.yaml.
Reported by @Spadav and @kotyKD.
Skills can now declare fallback_for_toolsets, fallback_for_tools,
requires_toolsets, and requires_tools in their SKILL.md frontmatter.
The system prompt builder filters skills automatically based on which
tools are available in the current session.
- Add _read_skill_conditions() to parse conditional frontmatter fields
- Add _skill_should_show() to evaluate conditions against available tools
- Update build_skills_system_prompt() to accept and apply tool availability
- Pass valid_tool_names and available toolsets from run_agent.py
- Backward compatible: skills without conditions always show; calling
build_skills_system_prompt() with no args preserves existing behavior
Closes#539
Implement send_document() and send_video() overrides in TelegramAdapter
so the agent can deliver files (PDFs, CSVs, docs, etc.) and videos as
native Telegram attachments instead of just printing the file path as
text.
The base adapter already routes MEDIA:<path> tags by extension — audio
goes to send_voice(), images to send_image_file(), and everything else
falls through to send_document(). But TelegramAdapter didn't override
send_document() or send_video(), so those fell back to plain text.
Now when the agent includes MEDIA:/path/to/report.pdf in its response,
users get a proper downloadable file attachment in Telegram.
Features:
- send_document: sends files via bot.send_document with display name,
caption (truncated to 1024), and reply_to support
- send_video: sends videos via bot.send_video with inline playback
- Both fall back to base class text if the Telegram API call fails
- 10 new tests covering success, custom filename, file-not-found,
not-connected, caption truncation, API error fallback, and reply_to
Requested by @TigerHixTang on Twitter.
- Register no-op app_mention event handler to suppress Bolt 404 errors.
The 'message' handler already processes @mentions in channels, so
app_mention is acknowledged without duplicate processing.
- Add send_document() for native file attachments (PDFs, CSVs, etc.)
via files_upload_v2, matching the pattern from Telegram PR #779.
- Add send_video() for native video uploads via files_upload_v2.
- Handle incoming document attachments from users: download, cache,
and inject text content for .txt/.md files (capped at 100KB),
following the same pattern as the Telegram adapter.
- Add _download_slack_file_bytes() helper for raw byte downloads.
- Add 24 new tests covering all new functionality.
Fixes the unhandled app_mention events reported in gateway logs.
Closes#643
Changes:
- /personality none|default|neutral — clears system prompt overlay
- Custom personalities in config.yaml support dict format with:
name, description, system_prompt, tone, style directives
- Backwards compatible — existing string format still works
- CLI + gateway both updated
- 18 tests covering none/default/neutral, dict format, string format,
list display, save to config
Add MCP sampling/createMessage capability via SamplingHandler class.
Text-only sampling + tool use in sampling with governance (rate limits,
model whitelist, token caps, tool loop limits). Per-server audit metrics.
Based on concept from PR #366 by eren-karakus0. Restructured as class-based
design with bug fixes and tests using real MCP SDK types.
50 new tests, 2600 total passing.
_FakeReadResult and _FakeSearchResult now expose the attributes
that read_file_tool/search_tool access after the redact_sensitive_text
integration from main.
Combine read/search loop detection with main's redact_sensitive_text
and truncation hint features. Add tracker reset to TestSearchHints
to prevent cross-test state leakage.
Add configurable bot message filtering via DISCORD_ALLOW_BOTS env var:
- 'none' (default): Ignore all other bot messages — matches previous
behavior where only our own bot was filtered, but now ALL bots are
filtered by default for cleaner channels
- 'mentions': Accept bot messages only when they @mention our bot —
useful for bot-to-bot workflows triggered by mentions
- 'all': Accept all bot messages — for setups where bots need to
interact freely
Previously, we only ignored our own bot's messages, allowing all other
bots through. This could cause noisy loops in channels with multiple bots.
8 new tests covering all filter modes and edge cases.
Inspired by openclaw v2026.3.7 Discord allowBots: 'mentions' config.
Enforce owner-only permissions on files and directories that contain
secrets or sensitive data:
- cron/jobs.py: jobs.json (0600), cron dirs (0700), job output files (0600)
- hermes_cli/config.py: config.yaml (0600), .env (0600), ~/.hermes/* dirs (0700)
- cli.py: config.yaml via save_config_value (0600)
All chmod calls use try/except for Windows compatibility.
Includes _secure_file() and _secure_dir() helpers with graceful fallback.
8 new tests verify permissions on all file types.
Inspired by openclaw v2026.3.7 file permission enforcement.
Two changes to prevent unnecessary Anthropic prompt cache misses in the
gateway, where a fresh AIAgent is created per user message:
1. Reuse stored system prompt for continuing sessions:
When conversation_history is non-empty, load the system prompt from
the session DB instead of rebuilding from disk. The model already has
updated memory in its conversation history (it wrote it!), so
re-reading memory from disk produces a different system prompt that
breaks the cache prefix.
2. Stabilize Honcho context per session:
- Only prefetch Honcho context on the first turn (empty history)
- Bake Honcho context into the cached system prompt and store to DB
- Remove the per-turn Honcho injection from the API call loop
This ensures the system message is identical across all turns in a
session. Previously, re-fetching Honcho could return different context
on each turn, changing the system message and invalidating the cache.
Both changes preserve the existing behavior for compression (which
invalidates the prompt and rebuilds from scratch) and for the CLI
(where the same AIAgent persists and the cached prompt is already
stable across turns).
Tests: 2556 passed (6 new)
Terminal output was already redacted via redact_sensitive_text() but
read_file and search_files returned raw content. Now both tools
redact secrets before returning results to the LLM.
Based on PR #372 by @teyrebaz33 (closes#363) — applied manually
due to branch conflicts with the current codebase.
The summary message was always injected as 'user' role, which causes
consecutive user messages when the last preserved head message is also
'user'. Some APIs reject this (400 error), and it produces malformed
training data.
Fix: check the role of the last head message and pick the opposite role
for the summary — 'user' after assistant/tool, 'assistant' after user.
Based on PR #328 by johnh4098. Closes#328.
Split fallback provider handling into two clean registries:
_FALLBACK_API_KEY_PROVIDERS — env-var-based (openrouter, zai, kimi, minimax)
_FALLBACK_OAUTH_PROVIDERS — OAuth-based (openai-codex, nous)
New _resolve_fallback_credentials() method handles all three cases
(OAuth, API key, custom endpoint) and returns a uniform (key, url, mode)
tuple. _try_activate_fallback() is now just validation + client build.
Adds Nous Portal as a fallback provider — uses the same OAuth flow
as the primary provider (hermes login), returns chat_completions mode.
OAuth providers get credential refresh for free: the existing 401
retry handlers (_try_refresh_codex/nous_client_credentials) check
self.provider, which is set correctly after fallback activation.
4 new tests (nous activation, nous no-login, codex retained).
27 total fallback tests passing, 2548 full suite.
Codex OAuth uses a different auth flow (OAuth tokens, not env vars)
and a different API mode (codex_responses, not chat_completions).
The fallback now handles this specially:
- Resolves credentials via resolve_codex_runtime_credentials()
- Sets api_mode to codex_responses
- Fails gracefully if no Codex OAuth session exists
Also added to the commented-out config.yaml example.
2 new tests (codex activation + graceful failure).
The gateway had a SEPARATE compression system ('session hygiene')
with hardcoded thresholds (100k tokens / 200 messages) that were
completely disconnected from the model's context length and the
user's compression config in config.yaml. This caused premature
auto-compression on Telegram/Discord — triggering at ~60k tokens
(from the 200-message threshold) or inconsistent token counts.
Changes:
- Gateway hygiene now reads model name from config.yaml and uses
get_model_context_length() to derive the actual context limit
- Compression threshold comes from compression.threshold in
config.yaml (default 0.85), same as the agent's ContextCompressor
- Removed the message-count-based trigger (was redundant and caused
false positives in tool-heavy sessions)
- Removed the undocumented session_hygiene config section — the
standard compression.* config now controls everything
- Env var overrides (CONTEXT_COMPRESSION_THRESHOLD,
CONTEXT_COMPRESSION_ENABLED) are respected
- Warn threshold is now 95% of model context (was hardcoded 200k)
- Updated tests to verify model-aware thresholds, scaling across
models, and that message count alone no longer triggers compression
For claude-opus-4.6 (200k context) at 85% threshold: gateway
hygiene now triggers at 170k tokens instead of the old 100k.
New browser capabilities and a built-in skill for agent-driven web QA.
## New tool: browser_console
Returns console messages (log/warn/error/info) AND uncaught JavaScript
exceptions in a single call. Uses agent-browser's 'console' and 'errors'
commands through the existing session plumbing. Supports --clear to reset
buffers. Verified working in both local and Browserbase cloud modes.
## Enhanced tool: browser_vision(annotate=True)
New boolean parameter on browser_vision. When true, agent-browser overlays
numbered [N] labels on interactive elements — each [N] maps to ref @eN.
Annotation data (element name, role, bounding box) returned alongside the
vision analysis. Useful for QA reports and spatial reasoning.
## Config: browser.record_sessions
Auto-record browser sessions as WebM video files when enabled:
- Starts recording on first browser_navigate
- Stops and saves on browser_close
- Saves to ~/.hermes/browser_recordings/
- Works in both local and cloud modes (verified)
- Disabled by default
## Built-in skill: dogfood
Systematic exploratory QA testing for web applications. Teaches the agent
a 5-phase workflow:
1. Plan — accept URL, create output dirs, set scope
2. Explore — systematic crawl with annotated screenshots
3. Collect Evidence — screenshots, console errors, JS exceptions
4. Categorize — severity (Critical/High/Medium/Low) and category
(Functional/Visual/Accessibility/Console/UX/Content)
5. Report — structured markdown with per-issue evidence
Includes:
- skills/dogfood/SKILL.md — full workflow instructions
- skills/dogfood/references/issue-taxonomy.md — severity/category defs
- skills/dogfood/templates/dogfood-report-template.md — report template
## Tests
21 new tests covering:
- browser_console message/error parsing, clear flag, empty/failed states
- browser_console schema registration
- browser_vision annotate schema and flag passing
- record_sessions config defaults and recording lifecycle
- Dogfood skill file existence and content validation
Addresses #315.
Remove hallucinated providers (openai, deepseek, together, groq,
fireworks, mistral, gemini, nous) from the fallback provider map.
These don't exist in hermes-agent's provider system.
The real supported providers for fallback are:
openrouter (OPENROUTER_API_KEY)
zai (ZAI_API_KEY)
kimi-coding (KIMI_API_KEY)
minimax (MINIMAX_API_KEY)
minimax-cn (MINIMAX_CN_API_KEY)
For any other OpenAI-compatible endpoint, users can use the
base_url + api_key_env overrides in the config.
Also adds Kimi User-Agent header for kimi fallback (matching
the main provider system).
When the primary model/provider fails after retries (rate limit, overload,
auth errors, connection failures), Hermes automatically switches to a
configured fallback model for the remainder of the session.
Config (in ~/.hermes/config.yaml):
fallback_model:
provider: openrouter
model: anthropic/claude-sonnet-4
Supports all major providers: OpenRouter, OpenAI, Nous, DeepSeek, Together,
Groq, Fireworks, Mistral, Gemini — plus custom endpoints via base_url and
api_key_env overrides.
Design principles:
- Dead simple: one fallback model, not a chain
- One-shot: switches once, doesn't ping-pong back
- Zero new dependencies: uses existing OpenAI client
- Minimal code: ~100 lines in run_agent.py, ~5 lines in cli.py/gateway
- Three trigger points: max retries exhausted, non-retryable client errors,
and invalid response exhaustion
Does NOT trigger on context overflow or payload-too-large errors (those
are handled by the existing compression system).
Addresses #737.
25 new tests, 2492 total passing.
Complete Signal adapter using signal-cli daemon HTTP API.
Based on PR #268 by ibhagwan, rebuilt on current main with bug fixes.
Architecture:
- SSE streaming for inbound messages with exponential backoff (2s→60s)
- JSON-RPC 2.0 for outbound (send, typing, attachments, contacts)
- Health monitor detects stale SSE connections (120s threshold)
- Phone number redaction in all logs and global redact.py
Features:
- DM and group message support with separate access policies
- DM policies: pairing (default), allowlist, open
- Group policies: disabled (default), allowlist, open
- Attachment download with magic-byte type detection
- Typing indicators (8s refresh interval)
- 100MB attachment size limit, 8000 char message limit
- E.164 phone + UUID allowlist support
Integration:
- Platform.SIGNAL enum in gateway/config.py
- Signal in _is_user_authorized() allowlist maps (gateway/run.py)
- Adapter factory in _create_adapter() (gateway/run.py)
- user_id_alt/chat_id_alt fields in SessionSource for UUIDs
- send_message tool support via httpx JSON-RPC (not aiohttp)
- Interactive setup wizard in 'hermes gateway setup'
- Connectivity testing during setup (pings /api/v1/check)
- signal-cli detection and install guidance
Bug fixes from PR #268:
- Timestamp reads from envelope_data (not outer wrapper)
- Uses httpx consistently (not aiohttp in send_message tool)
- SIGNAL_DEBUG scoped to signal logger (not root)
- extract_images regex NOT modified (preserves group numbering)
- pairing.py NOT modified (no cross-platform side effects)
- No dual authorization (adapter defers to run.py for user auth)
- Wildcard uses set membership ('*' in set, not list equality)
- .zip default for PK magic bytes (not .docx)
No new Python dependencies — uses httpx (already core).
External requirement: signal-cli daemon (user-installed).
Tests: 30 new tests covering config, init, helpers, session source,
phone redaction, authorization, and send_message integration.
Co-authored-by: ibhagwan <ibhagwan@users.noreply.github.com>
The gateway had a SEPARATE compression system ('session hygiene')
with hardcoded thresholds (100k tokens / 200 messages) that were
completely disconnected from the model's context length and the
user's compression config in config.yaml. This caused premature
auto-compression on Telegram/Discord — triggering at ~60k tokens
(from the 200-message threshold) or inconsistent token counts.
Changes:
- Gateway hygiene now reads model name from config.yaml and uses
get_model_context_length() to derive the actual context limit
- Compression threshold comes from compression.threshold in
config.yaml (default 0.85), same as the agent's ContextCompressor
- Removed the message-count-based trigger (was redundant and caused
false positives in tool-heavy sessions)
- Removed the undocumented session_hygiene config section — the
standard compression.* config now controls everything
- Env var overrides (CONTEXT_COMPRESSION_THRESHOLD,
CONTEXT_COMPRESSION_ENABLED) are respected
- Warn threshold is now 95% of model context (was hardcoded 200k)
- Updated tests to verify model-aware thresholds, scaling across
models, and that message count alone no longer triggers compression
For claude-opus-4.6 (200k context) at 85% threshold: gateway
hygiene now triggers at 170k tokens instead of the old 100k.
The 'openai' provider was redundant — using OPENAI_BASE_URL +
OPENAI_API_KEY with provider: 'main' already covers direct OpenAI API.
Provider options are now: auto, openrouter, nous, codex, main.
- Removed _try_openai(), _OPENAI_AUX_MODEL, _OPENAI_BASE_URL
- Replaced openai tests with codex provider tests
- Updated all docs to remove 'openai' option and clarify 'main'
- 'main' description now explicitly mentions it works with OpenAI API,
local models, and any OpenAI-compatible endpoint
Tests: 2467 passed.
The Codex Responses API (chatgpt.com/backend-api/codex) supports
vision via gpt-5.3-codex. This was verified with real API calls
using image analysis.
Changes to _CodexCompletionsAdapter:
- Added _convert_content_for_responses() to translate chat.completions
multimodal format to Responses API format:
- {type: 'text'} → {type: 'input_text'}
- {type: 'image_url', image_url: {url: '...'}} → {type: 'input_image', image_url: '...'}
- Fixed: removed 'stream' from resp_kwargs (responses.stream() handles it)
- Fixed: removed max_output_tokens and temperature (Codex endpoint rejects them)
Provider changes:
- Added 'codex' as explicit auxiliary provider option
- Vision auto-fallback now includes Codex (OpenRouter → Nous → Codex)
since gpt-5.3-codex supports multimodal input
- Updated docs with Codex OAuth examples
Tested with real Codex OAuth token + ~/.hermes/image2.png — confirmed
working end-to-end through the full adapter pipeline.
Tests: 2459 passed.
The Codex model normalization was rejecting any model without 'codex'
in its name, forcing a fallback to gpt-5.3-codex. This blocked models
like gpt-5.4 that the Codex API actually supports.
The fix simplifies _normalize_model_for_provider() to two operations:
1. Strip provider prefixes (API needs bare slugs)
2. Replace the *untouched default* model with a Codex-compatible one
If the user explicitly chose a model — any model — we trust them and
let the API be the judge. No allowlists, no slug checks.
Also removes the 'codex not in slug' filter from _read_cache_models()
so the local cache preserves all API-available models.
Inspired by OpenClaw's approach which explicitly lists non-codex models
(gpt-5.4, gpt-5.2) as valid Codex models.
Users can now set provider: "openai" for auxiliary tasks (vision, web
extract, compression) to use OpenAI's API directly with their
OPENAI_API_KEY. This hits api.openai.com/v1 with gpt-4o-mini as the
default model — supports vision since GPT-4o handles image input.
Provider options are now: auto, openrouter, nous, openai, main.
Changes:
- agent/auxiliary_client.py: added _try_openai(), "openai" case in
_resolve_forced_provider(), updated auxiliary_max_tokens_param()
to use max_completion_tokens for OpenAI
- Updated docs: cli-config.yaml.example, AGENTS.md, and user-facing
configuration.md with Common Setups section showing OpenAI,
OpenRouter, and local model examples
- 3 new tests for OpenAI provider resolution
Tests: 2459 passed (was 2429).
Improvements on top of PR #606 (auxiliary model configuration):
1. Gateway bridge: Added auxiliary.* and compression.summary_provider
config bridging to gateway/run.py so config.yaml settings work from
messaging platforms (not just CLI). Matches the pattern in cli.py.
2. Vision auto-fallback safety: In auto mode, vision now only tries
OpenRouter + Nous Portal (known multimodal-capable providers).
Custom endpoints, Codex, and API-key providers are skipped to avoid
confusing errors from providers that don't support vision input.
Explicit provider override (AUXILIARY_VISION_PROVIDER=main) still
allows using any provider.
3. Comprehensive tests (46 new):
- _get_auxiliary_provider env var resolution (8 tests)
- _resolve_forced_provider with all provider types (8 tests)
- Per-task provider routing integration (4 tests)
- Vision auto-fallback safety (7 tests)
- Config bridging logic (11 tests)
- Gateway/CLI bridge parity (2 tests)
- Vision model override via env var (2 tests)
- DEFAULT_CONFIG shape validation (4 tests)
4. Docs: Added auxiliary_client.py to AGENTS.md project structure.
Updated module docstring with separate text/vision resolution chains.
Tests: 2429 passed (was 2383).
- Added support for auxiliary model overrides in the configuration, allowing users to specify providers and models for vision and web extraction tasks.
- Updated the CLI configuration example to include new auxiliary model settings.
- Enhanced the environment variable mapping in the CLI to accommodate auxiliary model configurations.
- Improved the resolution logic for auxiliary clients to support task-specific provider overrides.
- Updated relevant documentation and comments for clarity on the new features and their usage.
Add contextual [Hint: ...] suffixes to tool results where they save
real iterations:
- patch (no match): suggests read_file/search_files to verify content
before retrying — addresses the common pattern where the agent retries
with stale old_string instead of re-reading the file.
- search_files (truncated): provides explicit next offset and suggests
narrowing the search — clearer than relying on total_count inference.
Other hints proposed in #722 (terminal, web_search, web_extract,
browser_snapshot, search zero-results, search content-matches) were
evaluated and found to be low-value: either already covered by existing
mechanisms (read_file pagination, similar-files, schema descriptions)
or guidance the agent already follows from its own reasoning.
5 new tests covering hint presence/absence for both tools.
When resuming a session via --continue or --resume, show a compact recap
of the previous conversation inside a Rich panel before the input prompt.
This gives users immediate visual context about what was discussed.
Changes:
- Add _preload_resumed_session() to load session history early (in run(),
before banner) so _init_agent() doesn't need a separate DB round-trip
- Add _display_resumed_history() that renders a formatted recap panel:
* User messages shown with gold bullet (truncated at 300 chars)
* Assistant responses shown with green diamond (truncated at 200 chars / 3 lines)
* Tool calls collapsed to count + tool names
* System messages and tool results hidden
* <REASONING_SCRATCHPAD> blocks stripped from display
* Pure-reasoning messages (no visible output) skipped entirely
* Capped at last 10 exchanges with 'N earlier messages' indicator
* Dim/muted styling distinguishes recap from active conversation
- Add display.resume_display config option: 'full' (default) or 'minimal'
- Store resume_display as instance variable (like compact) for testability
- 27 new tests covering all display scenarios, config, and edge cases
Closes#719
NOUS_API_KEY is unused — vision tools use OPENROUTER_API_KEY or Nous
Portal OAuth (auth.json), and MoA tools use OPENROUTER_API_KEY.
Removed from:
- hermes_cli/config.py: api_keys allowlist for config set routing
- .env.example: example env file entry and comment
- tests/hermes_cli/test_set_config_value.py: parametrize test data
- tests/integration/test_web_tools.py: updated comments and log
messages to reference 'auxiliary LLM provider' instead of NOUS_API_KEY
No HECATE references found in codebase (already cleaned up).
Add `hermes sessions browse` — a curses-based interactive session picker
with live type-to-search filtering, arrow key navigation, and seamless
session resume via Enter.
Features:
- Arrow keys to navigate, Enter to select and resume, Esc/q to quit
- Type characters to live-filter sessions by title, preview, source, or ID
- Backspace to edit filter, first Esc clears filter, second Esc exits
- Adaptive column layout (title/preview, last active, source, ID)
- Scrolling support for long session lists
- --source flag to filter by platform (cli, telegram, discord, etc.)
- --limit flag to control how many sessions to load (default: 50)
- Windows fallback: numbered list with input prompt
- After selection, seamlessly execs into `hermes --resume <id>`
Design decisions:
- Separate subcommand (not a flag on -c) — preserves `hermes -c` as-is
for instant most-recent-session resume
- Uses curses (not simple_term_menu) per Known Pitfalls to avoid the
arrow-key ghost-duplication rendering bug in tmux/iTerm
- Follows existing curses pattern from hermes_cli/tools_config.py
Also fixes: removed redundant `import os` inside cmd_sessions stats
block that shadowed the module-level import (would cause UnboundLocalError
if browse action was taken in the same function).
Tests: 33 new tests covering curses picker, fallback mode, filtering,
navigation, edge cases, and argument parser registration.
Verifies that setup.py imports the correct function name
(get_codex_model_ids) from codex_models.py. This would have caught
the ImportError bug before it reached users.
Source code (hermes_cli/clipboard.py):
- _convert_to_png() lost the file when both Pillow and ImageMagick were
unavailable: path.rename(tmp) moved the file to .bmp, then subprocess.run
raised FileNotFoundError, but the file was never renamed back. The final
fallback 'return path.exists()' returned False.
- Fix: restore the original file in both except handlers by renaming tmp
back to path when the original is missing.
Test (tests/tools/test_clipboard.py):
- test_file_still_usable_when_no_converter expected 'from PIL import Image'
to raise an Exception, but Pillow is installed so pytest.raises fired
'DID NOT RAISE'. The test also never called _convert_to_png().
- Fix: properly mock PIL unavailability via patch.dict(sys.modules),
actually call _convert_to_png(), and assert the correct result.
Messaging users can now switch back to previously-named sessions:
- /resume My Project — resolves the title (with auto-lineage) and
restores that session's conversation history
- /resume (no args) — lists recent titled sessions to choose from
Adds SessionStore.switch_session() which ends the current session and
points the session entry at the target session ID so the old transcript
is loaded on the next message. Running agents are cleared on switch.
Completes the session naming feature from PR #720 for gateway users.
8 new tests covering: name resolution, lineage auto-latest, already-on-
session check, nonexistent names, agent cleanup, no-DB fallback, and
listing titled sessions.
When _ensure_runtime_credentials() resolves the provider to openai-codex,
check if the active model is Codex-compatible. If not (e.g. the default
anthropic/claude-opus-4.6), swap it for the best available Codex model.
Also strips provider prefixes the Codex API rejects (openai/gpt-5.3-codex
→ gpt-5.3-codex).
Adds _model_is_default flag so warnings are only shown when the user
explicitly chose an incompatible model (not when it's the config default).
Fixes#651.
Co-inspired-by: stablegenius49 (PR #661)
Co-inspired-by: teyrebaz33 (PR #696)
Previously, search_files would silently return 0 results when the
search path didn't exist (e.g., /root/.hermes/... when HOME is
/home/user). The path was passed to rg/grep/find which would fail
silently, and the empty stdout was parsed as 'no matches found'.
Changes:
- Add path existence check at the top of search() using test -e.
Returns SearchResult with a clear error message when path doesn't exist.
- Add exit code 2 checks in _search_with_rg() and _search_with_grep()
as secondary safety net for other error types (bad regex, permissions).
- Add 4 new tests covering: nonexistent path (content mode), nonexistent
path (files mode), existing path proceeds normally, rg error exit code.
Tests: 37 → 41 in test_file_operations.py, full suite 2330 passed.
Adds a fun alias for skipping all dangerous command approval prompts.
When passed, sets HERMES_YOLO_MODE=1 which causes check_dangerous_command()
to auto-approve everything.
Available on both top-level and chat subcommand:
hermes --fuck-it-ship-it
hermes chat --fuck-it-ship-it
Includes 5 tests covering normal blocking, yolo bypass, all patterns,
and edge cases (empty string env var).
- Empty string titles normalized to None (prevents uncaught IntegrityError
when two sessions both get empty-string titles via the unique index)
- Escape SQL LIKE wildcards (%, _) in resolve_session_by_title and
get_next_title_in_lineage to prevent false matches on titles like
'test_project' matching 'testXproject #2'
- Optimize list_sessions_rich from N+2 queries to a single query with
correlated subqueries (preview + last_active computed in SQL)
- Add /title slash command to gateway (Telegram, Discord, Slack, WhatsApp)
with set and show modes, uniqueness conflict handling
- Add /title to gateway /help text and _known_commands
- 12 new tests: empty string normalization, multi-empty-title safety,
SQL wildcard edge cases, gateway /title set/show/conflict/cross-platform
Two 'except Exception: pass' blocks silently hide failures:
- mirror_to_session failure: user's message never gets mirrored, no trace
- config.yaml parse failure: wrong model used silently
Replace with logger.warning so failures are visible in logs.
Completed/cancelled items are now filtered from format_for_injection()
output. Update the existing test to verify active items appear and
completed items are excluded.
- Block file reads after 3+ re-reads of same region (no content returned)
- Track search_files calls and block repeated identical searches
- Filter completed/cancelled todos from post-compression injection
to prevent agent from re-doing finished work
- Add 10 new tests covering all three fixes
When context compression summarizes conversation history, the agent
loses track of which files it already read and re-reads them in a loop.
Users report the agent reading the same files endlessly without writing.
Root cause: context compression is lossy — file contents and read history
are lost in the summary. After compression, the model thinks it hasn't
examined the files yet and reads them again.
Fix (two-part):
1. Track file reads per task in file_tools.py. When the same file region
is read again, include a _warning in the response telling the model
to stop re-reading and use existing information.
2. After context compression, inject a structured message listing all
files already read in the session with explicit "do NOT re-read"
instruction, preserving read history across compression boundaries.
Adds 16 tests covering warning detection, task isolation, summary
accuracy, tracker cleanup, and compression history injection.
Call ensure_hub_dirs() at the start of hermes skills list so the\nSkills Hub directory structure is created before reading hub\nmetadata.\n\nAdd a regression test covering the empty-home path where\ndoctor recommends running the list command.\n\nRefs: #703
Images pasted in the CLI were embedded as raw base64 image_url content
parts in the conversation history, which only works with vision-capable
models. If the main model (e.g. Nous API) doesn't support vision, this
breaks the request and poisons all subsequent messages.
Now the CLI uses the same approach as the messaging gateway: images are
pre-processed through the auxiliary vision model (Gemini Flash via
OpenRouter or Nous Portal) and converted to text descriptions. The
local file path is included so the agent can re-examine via
vision_analyze if needed. Works with any model.
Fixes#638.
User messaging improvements:
- Rejection: '(>_<) Error: not a valid model' instead of '(^_^) Warning: Error:'
- Rejection: shows 'Model unchanged' + tip about /model and /provider
- Session-only: explains 'this session only' with reason and 'will revert on restart'
- Saved: clear '(saved to config)' confirmation
Docs updated:
- cli-commands.md, cli.md, messaging/index.md: /model now shows
provider:model syntax, /provider command added to tables
Test fixes: deduplicated test names, assertions match new messages.
/provider command (CLI + gateway):
Shows all providers with auth status (✓/✗), aliases, and active marker.
Users can now discover what provider names work with provider:model syntax.
Gateway bugs fixed:
- Config was saved even when validation.persist=False (told user 'session
only' but actually persisted the unvalidated model)
- HERMES_INFERENCE_PROVIDER env var not set on provider switch, causing
the switch to be silently overridden if that env var was already set
parse_model_input hardened:
- Colon only treated as provider delimiter if left side is a recognized
provider name or alias. 'anthropic/claude-3.5-sonnet:beta' now passes
through as a model name instead of trying provider='anthropic/claude-3.5-sonnet'.
- HTTP URLs, random colons no longer misinterpreted.
56 tests passing across model validation, CLI commands, and integration.
Add provider:model syntax to /model command for runtime provider switching:
/model zai:glm-5 → switch to Z.AI provider with glm-5
/model nous:hermes-3 → switch to Nous Portal with hermes-3
/model openrouter:anthropic/claude-sonnet-4.5 → explicit OpenRouter
When switching providers, credentials are resolved via resolve_runtime_provider
and validated before committing. Both model and provider are saved to config.
Provider aliases work (glm: → zai, kimi: → kimi-coding, etc.).
Enhanced /model (no args) display now shows:
- Current model and provider
- Curated model list for the current provider with ← marker
- Usage examples including provider:model syntax
39 tests covering parse_model_input, curated_models_for_provider,
provider switching (success + credential failure), and display output.
The 200 lines of prompt_toolkit/rich/fire stubs added in PR #650 were
guarded by 'if module in sys.modules: return' and never activated since
those dependencies are always installed. Removed to keep the test file
lean. Also removed unused MagicMock and pytest imports.
Not all providers require 'provider/model' format. Removing the rigid
format check lets the live API probe handle all validation uniformly.
If someone types 'gpt-5.4' on OpenRouter, the probe won't find it and
will suggest 'openai/gpt-5.4' — better UX than a format rejection.
Replace the static catalog-based model validation with a live API probe.
The /model command now hits the provider's /models endpoint to check if
the requested model actually exists:
- Model found in API → accepted + saved to config
- Model NOT found in API → rejected with 'Error: not a valid model'
and fuzzy-match suggestions from the live model list
- API unreachable → graceful fallback to hardcoded catalog (session-only
for unrecognized models)
- Format errors (empty, spaces, missing '/') still caught instantly
without a network call
The API probe takes ~0.2s for OpenRouter (346 models) and works with any
OpenAI-compatible endpoint (Ollama, vLLM, custom, etc.).
32 tests covering all paths: format checks, API found, API not found,
API unreachable fallback, CLI integration.
- Wrap validate_requested_model in try/except so /model doesn't crash
if validation itself fails (falls back to old accept+save behavior)
- Remove unnecessary sys.path.insert from both test files
- Expand test_model_validation.py: 4 → 23 tests covering normalize_provider,
provider_model_ids, empty/whitespace/spaces rejection, OpenRouter format
validation, custom endpoints, nous provider, provider aliases, unknown
providers, fuzzy suggestions
- Expand test_cli_model_command.py: 2 → 5 tests adding known-model save,
validation crash fallback, and /model with no argument
Skills can now declare runtime prerequisites (env vars, CLI binaries) via
YAML frontmatter. Skills with unmet prerequisites are excluded from the
system prompt so the agent never claims capabilities it can't deliver, and
skill_view() warns the agent about what's missing.
Three layers of defense:
- build_skills_system_prompt() filters out unavailable skills
- _find_all_skills() flags unmet prerequisites in metadata
- skill_view() returns prerequisites_warning with actionable details
Tagged 12 bundled skills that have hard runtime dependencies:
gif-search (TENOR_API_KEY), notion (NOTION_API_KEY), himalaya, imessage,
apple-notes, apple-reminders, openhue, duckduckgo-search, codebase-inspection,
blogwatcher, songsee, mcporter.
Closes#658Fixes#630
Removed the hard block on base_url containing 'api.anthropic.com'.
Anthropic now offers an OpenAI-compatible /chat/completions endpoint,
so blocking their URL prevents legitimate use. If the endpoint isn't
compatible, the API call will fail with a proper error anyway.
Removed from: run_agent.py, mini_swe_runner.py
Updated test to verify Anthropic URLs are accepted.
browser_vision now saves screenshots persistently to ~/.hermes/browser_screenshots/
and returns the screenshot_path in its JSON response. The model can include
MEDIA:<path> in its response to share screenshots as native photos.
Changes:
- browser_tool.py: Save screenshots persistently, return screenshot_path,
auto-cleanup files older than 24 hours, mkdir moved inside try/except
- telegram.py: Add send_image_file() — sends local images via bot.send_photo()
- discord.py: Add send_image_file() — sends local images via discord.File
- slack.py: Add send_image_file() — sends local images via files_upload_v2()
(WhatsApp already had send_image_file — no changes needed)
- prompt_builder.py: Updated Telegram hint to list image extensions,
added Discord and Slack MEDIA: platform hints
- browser.md: Document screenshot sharing and 24h cleanup
- send_file_integration_map.md: Updated to reflect send_image_file is now
implemented on Telegram/Discord/Slack
- test_send_image_file.py: 19 tests covering MEDIA: .png extraction,
send_image_file on all platforms, and screenshot cleanup
Partially addresses #466 (Phase 0: platform adapter gaps for send_image_file).
Authored by christomitov. Auto-detects sk-kimi- key prefix and routes
to api.kimi.com/coding/v1. Adds User-Agent header for Kimi Code API
compatibility. Legacy Moonshot keys continue to work unchanged.
Critical fixes:
- Add --worktree/-w to hermes_cli/main.py argparse (both chat
subcommand and top-level parser) so 'hermes -w' works via the
actual CLI entry point, not just 'python cli.py -w'
- Pass worktree flag through cmd_chat() kwargs to cli_main()
- Handle worktree attr in bare 'hermes' and --resume/--continue paths
Bug fixes in cli.py:
- Skip worktree creation for --list-tools/--list-toolsets (wasteful)
- Wrap git worktree subprocess.run in try/except (crash on timeout)
- Add stale worktree pruning on startup (_prune_stale_worktrees):
removes clean worktrees older than 24h left by crashed/killed sessions
Documentation updates:
- AGENTS.md: add --worktree to CLI commands table
- cli-config.yaml.example: add worktree config section
- website/docs/reference/cli-commands.md: add to core commands
- website/docs/user-guide/cli.md: add usage examples
- website/docs/user-guide/configuration.md: add config docs
Test improvements (17 → 31 tests):
- Stale worktree pruning (prune old clean, keep recent, keep dirty)
- Directory symlink via .worktreeinclude
- Edge cases (no commits, not a repo, pre-existing .worktrees/)
- CLI flag/config OR logic
- TERMINAL_CWD integration
- System prompt injection format
Add a --worktree (-w) flag to the hermes CLI that creates an isolated
git worktree for the session. This allows running multiple hermes-agent
instances concurrently on the same repo without file collisions.
How it works:
- On startup with -w: detects git repo, creates .worktrees/<session>/
with its own branch (hermes/<session-id>), sets TERMINAL_CWD to it
- Each agent works in complete isolation — independent HEAD, index,
and working tree, shared git object store
- On exit: auto-removes worktree and branch if clean, warns and
keeps if there are uncommitted changes
- .worktreeinclude file support: list gitignored files (.env, .venv/)
to auto-copy/symlink into new worktrees
- .worktrees/ is auto-added to .gitignore
- Agent gets a system prompt note about the worktree context
- Config support: set worktree: true in config.yaml to always enable
Usage:
hermes -w # Interactive mode in worktree
hermes -w -q "Fix issue #123" # Single query in worktree
# Or in config.yaml:
worktree: true
Includes 17 tests covering: repo detection, worktree creation,
independence verification, cleanup (clean/dirty), .worktreeinclude,
.gitignore management, and 10 concurrent worktrees.
Closes#652
Long-lived gateway sessions can accumulate enough history that every new
message rehydrates an oversized transcript, causing repeated truncation
failures (finish_reason=length).
Add a session hygiene check in _handle_message that runs right after
loading the transcript and before invoking the agent:
1. Estimate message count and rough token count of the transcript
2. If above configurable thresholds (default: 200 msgs or 100K tokens),
auto-compress the transcript proactively
3. Notify the user about the compression with before/after stats
4. If still above warn threshold (default: 200K tokens) after
compression, suggest /reset
5. If compression fails on a dangerously large session, warn the user
to use /compress or /reset manually
Thresholds are configurable via config.yaml:
session_hygiene:
auto_compress_tokens: 100000
auto_compress_messages: 200
warn_tokens: 200000
This complements the agent's existing preflight compression (which
runs inside run_conversation) by catching pathological sessions at
the gateway layer before the agent is even created.
Includes 12 tests for threshold detection and token estimation.
Kimi Code (platform.kimi.ai) issues API keys prefixed sk-kimi- that require:
1. A different base URL: api.kimi.com/coding/v1 (not api.moonshot.ai/v1)
2. A User-Agent header identifying a recognized coding agent
Without this fix, sk-kimi- keys fail with 401 (wrong endpoint) or 403
('only available for Coding Agents') errors.
Changes:
- Auto-detect sk-kimi- key prefix and route to api.kimi.com/coding/v1
- Send User-Agent: KimiCLI/1.0 header for Kimi Code endpoints
- Legacy Moonshot keys (api.moonshot.ai) continue to work unchanged
- KIMI_BASE_URL env var override still takes priority over auto-detection
- Updated .env.example with correct docs and all endpoint options
- Fixed doctor.py health check for Kimi Code keys
Reference: https://github.com/MoonshotAI/kimi-cli (platforms.py)
Previously, when a session expired (idle/daily reset), the memory flush
ran synchronously inside get_or_create_session — blocking the user's
message for 10-60s while an LLM call saved memories.
Now a background watcher task (_session_expiry_watcher) runs every 5 min,
detects expired sessions, and flushes memories proactively in a thread
pool. By the time the user sends their next message, memories are
already saved and the response is immediate.
Changes:
- Add _is_session_expired(entry) to SessionStore — works from entry
alone without needing a SessionSource
- Add _pre_flushed_sessions set to track already-flushed sessions
- Remove sync _on_auto_reset callback from get_or_create_session
- Refactor flush into _flush_memories_for_session (sync worker) +
_async_flush_memories (thread pool wrapper)
- Add _session_expiry_watcher background task, started in start()
- Simplify /reset command to use shared fire-and-forget flush
- Add 10 tests for expiry detection, callback removal, tracking
Reduces token usage and latency for most tasks by defaulting to
medium reasoning effort instead of xhigh. Users can still override
via config or CLI flag. Updates code, tests, example config, and docs.
_convert_to_png() renamed the original file to .bmp before calling
ImageMagick convert, then unconditionally deleted the .bmp regardless
of whether convert succeeded. If convert failed, both files were gone.
- Only delete .bmp after confirmed successful conversion
- Restore original file on convert failure, timeout, or missing binary
- Add 3 tests covering failure, not-installed, and timeout scenarios
_make_cli() did not clear HERMES_MAX_ITERATIONS env var, so tests
failed in CI where the var was set externally. Also, default max_turns
changed from 60 to 90 in 0a82396 but tests were not updated.
- Clear HERMES_MAX_ITERATIONS in _make_cli() for proper isolation
- Add env_overrides parameter for tests that need specific env values
- Update hardcoded 60 assertions to 90 to match new default
- Simplify test_env_var_max_turns using env_overrides
When MarkdownV2 parsing fails, _strip_mdv2() removes escape backslashes
and bold markers (*text*) but missed italic markers (_text_). Users saw
raw underscores around italic text in the plaintext fallback.
- Add regex to strip _text_ italic markers in _strip_mdv2()
- Use word boundary lookaround to preserve snake_case identifiers
- Add tests for _strip_mdv2 covering italic, bold, snake_case, and edge cases
Add a 'platforms' field to SKILL.md frontmatter that restricts skills
to specific operating systems. Skills with platforms: [macos] only
appear in the system prompt, skills_list(), and slash commands on macOS.
Skills without the field load everywhere (backward compatible).
Implementation:
- skill_matches_platform() in tools/skills_tool.py — core filter
- Wired into all 3 discovery paths: prompt_builder.py, skills_tool.py,
skill_commands.py
- 28 new tests across 3 test files
New bundled Apple/macOS skills (all platforms: [macos]):
- imessage — Send/receive iMessages via imsg CLI
- apple-reminders — Manage Reminders via remindctl CLI
- apple-notes — Manage Notes via memo CLI
- findmy — Track devices/AirTags via AppleScript + screen capture
Docs updated: CONTRIBUTING.md, AGENTS.md, creating-skills.md,
skills.md (user guide)
Authored by areu01or00. Adds timezone support via hermes_time.now() helper
with IANA timezone resolution (HERMES_TIMEZONE env → config.yaml → server-local).
Updates system prompt timestamp, cron scheduling, and execute_code sandbox TZ
injection. Includes config migration (v4→v5) and comprehensive test coverage.
Introduces a new OpenClaw-to-Hermes migration skill with a Python
helper script that handles importing SOUL.md, memories, user profiles,
messaging settings, command allowlists, skills, TTS assets, and
workspace instructions.
Supports two migration presets (user-data / full), three skill conflict
modes (skip / overwrite / rename), overflow file export for entries that
exceed character limits, and granular include/exclude option filtering.
Includes detailed SKILL.md agent instructions covering the clarify-tool
interaction protocol, decision-to-command mapping, post-run reporting
rules, and path resolution guidance.
Adds dynamic panel width calculation to CLI clarify/approval widgets so
panels adapt to content and terminal size.
Includes 7 new tests covering presets, include/exclude, conflict modes,
overflow exports, and skills_guard integration.
Adds 4 new direct API-key providers (zai, kimi-coding, minimax, minimax-cn)
to the inference provider system. All use standard OpenAI-compatible
chat/completions endpoints with Bearer token auth.
Core changes:
- auth.py: Extended ProviderConfig with api_key_env_vars and base_url_env_var
fields. Added providers to PROVIDER_REGISTRY. Added provider aliases
(glm, z-ai, zhipu, kimi, moonshot). Added auto-detection of API-key
providers in resolve_provider(). Added resolve_api_key_provider_credentials()
and get_api_key_provider_status() helpers.
- runtime_provider.py: Added generic API-key provider branch in
resolve_runtime_provider() — any provider with auth_type='api_key'
is automatically handled.
- main.py: Added providers to hermes model menu with generic
_model_flow_api_key_provider() flow. Updated _has_any_provider_configured()
to check all provider env vars. Updated argparse --provider choices.
- setup.py: Added providers to setup wizard with API key prompts and
curated model lists.
- config.py: Added env vars (GLM_API_KEY, KIMI_API_KEY, MINIMAX_API_KEY,
etc.) to OPTIONAL_ENV_VARS.
- status.py: Added API key display and provider status section.
- doctor.py: Added connectivity checks for each provider endpoint.
- cli.py: Updated provider docstrings.
Docs: Updated README.md, .env.example, cli-config.yaml.example,
cli-commands.md, environment-variables.md, configuration.md.
Tests: 50 new tests covering registry, aliases, resolution, auto-detection,
credential resolution, and runtime provider dispatch.
Inspired by PR #33 (numman-ali) which proposed a provider registry approach.
Credit to tars90percent (PR #473) and manuelschipper (PR #420) for related
provider improvements merged earlier in this changeset.
Two bugs fixed:
1. search_messages() crashes with OperationalError when user queries
contain FTS5 special characters (+, ", (, {, dangling AND/OR, etc).
Added _sanitize_fts5_query() to strip dangerous operators and a
fallback try-except for edge cases.
2. _append_to_sqlite() in mirror.py creates a new SessionDB per call
but never closes it, leaking SQLite connections. Added finally block
to ensure db.close() is always called.
API key selection is now base_url-aware: when the resolved base_url
targets OpenRouter, OPENROUTER_API_KEY takes priority (preserving the
#289 fix). When hitting any other endpoint (Z.ai, vLLM, custom, etc.),
OPENAI_API_KEY takes priority so the OpenRouter key doesn't leak.
Applied in both the runtime provider resolver (the real code path) and
the CLI initial default (for consistency).
Fixes#560.
_make_cli() now patches CLI_CONFIG with clean defaults so
test_cli_init tests don't depend on the developer's local config.yaml.
test_empty_dir_returns_empty now mocks Path.home() so it doesn't pick
up a global SOUL.md.
Credit to teyrebaz33 for identifying and fixing these in PR #557.
Fixes#555.
tool_call_count was inaccurate in two ways:
1. Under-counting: an assistant message with N parallel tool calls
(e.g. "kill the light and shut off the fan" = 2 ha_call_service)
only incremented tool_call_count by 1 instead of N.
2. Over-counting: tool response messages (role=tool) also incremented
tool_call_count, double-counting every tool interaction.
Combined: 2 parallel tool calls produced tool_call_count=3 (1 from
assistant + 2 from tool responses) instead of the correct value of 2.
Fix: only count from assistant messages with tool_calls, incrementing
by len(tool_calls) to handle parallel calls correctly. Tool response
messages no longer affect tool_call_count.
This impacts /insights and /usage accuracy for sessions with tool use.
Two bugs in sync_skills():
1. Failed copytree poisons manifest: when shutil.copytree fails (disk
full, permission error), the skill is still recorded in the manifest.
On the next sync, the skill appears as "in manifest but not on disk"
which is interpreted as "user deliberately deleted it" — the skill
is never retried. Fix: only write to manifest on successful copy.
2. Failed update destroys user copy: rmtree deletes the existing skill
directory before copytree runs. If copytree then fails, the user's
skill is gone with no way to recover. Fix: move to .bak before
copying, restore from backup if copytree fails.
Both bugs are proven by new regression tests that fail on the old code
and pass on the fix.
Upgrade skills_sync manifest to v2 format (name:origin_hash). The origin
hash records the MD5 of the bundled skill at the time it was last synced.
On update, the user's copy is compared against the origin hash:
- User copy == origin hash → unmodified → safe to update from bundled
- User copy != origin hash → user customized → skip (preserve changes)
v1 manifests (plain names) are auto-migrated: the user's current hash
becomes the baseline, so future syncs can detect modifications.
Output now shows user-modified skills:
~ whisper (user-modified, skipping)
27 tests covering all scenarios including v1→v2 migration, user
modification detection, update after migration, and origin hash tracking.
2009 tests pass.
- Restored 21 skills removed in commits 757d012 and 740dd92:
accelerate, audiocraft, code-review, faiss, flash-attention, gguf,
grpo-rl-training, guidance, llava, nemo-curator, obliteratus, peft,
pytorch-fsdp, pytorch-lightning, simpo, slime, stable-diffusion,
tensorrt-llm, torchtitan, trl-fine-tuning, whisper
- Rewrote sync_skills() with proper update semantics:
* New skills (not in manifest): copied to user dir
* Existing skills (in manifest + on disk): updated via hash comparison
* User-deleted skills (in manifest, not on disk): respected, not re-added
* Stale manifest entries (removed from bundled): cleaned from manifest
- Added sync_skills() to CLI startup (cmd_chat) and gateway startup
(start_gateway) — previously only ran during 'hermes update'
- Updated cmd_update output to show new/updated/cleaned counts
- Rewrote tests: 20 tests covering manifest CRUD, dir hashing, fresh
install, user deletion respect, update detection, stale cleanup, and
name collision handling
75 bundled skills total. 2002 tests pass.
Issues found and fixed during deep code path review:
1. CRITICAL: Prefix matching returned wrong prices for dated model names
- 'gpt-4o-mini-2024-07-18' matched gpt-4o ($2.50) instead of gpt-4o-mini ($0.15)
- Same for o3-mini→o3 (9x), gpt-4.1-mini→gpt-4.1 (5x), gpt-4.1-nano→gpt-4.1 (20x)
- Fix: use longest-match-wins strategy instead of first-match
- Removed dangerous key.startswith(bare) reverse matching
2. CRITICAL: Top Tools section was empty for CLI sessions
- run_agent.py doesn't set tool_name on tool response messages (pre-existing)
- Insights now also extracts tool names from tool_calls JSON on assistant
messages, which IS populated for all sessions
- Uses max() merge strategy to avoid double-counting between sources
3. SELECT * replaced with explicit column list
- Skips system_prompt and model_config blobs (can be thousands of chars)
- Reduces memory and I/O for large session counts
4. Sets in overview dict converted to sorted lists
- models_with_pricing / models_without_pricing were Python sets
- Sets aren't JSON-serializable — would crash json.dumps()
5. Negative duration guard
- end > start check prevents negative durations from clock drift
6. Model breakdown sort fallback
- When all tokens are 0, now sorts by session count instead of arbitrary order
7. Removed unused timedelta import
Added 6 new tests: dated model pricing (4), tool_calls JSON extraction,
JSON serialization safety. Total: 69 tests.
Custom OAI endpoints, self-hosted models, and local inference should NOT
show fabricated cost estimates. Changed default pricing from $3/$12 per
million tokens to $0/$0 for unrecognized models.
- Added _has_known_pricing() to distinguish commercial vs custom models
- Models with known pricing show $ amounts; unknown models show 'N/A'
- Overview shows asterisk + note when some models lack pricing data
- Gateway format adds '(excludes custom/self-hosted models)' note
- Added 7 new tests for custom model cost handling
Inspired by Claude Code's /insights, adapted for Hermes Agent's multi-platform
architecture. Analyzes session history from state.db to produce comprehensive
usage insights.
Features:
- Overview stats: sessions, messages, tokens, estimated cost, active time
- Model breakdown: per-model sessions, tokens, and cost estimation
- Platform breakdown: CLI vs Telegram vs Discord etc. (unique to Hermes)
- Tool usage ranking: most-used tools with percentages
- Activity patterns: day-of-week chart, peak hours, streaks
- Notable sessions: longest, most messages, most tokens, most tool calls
- Cost estimation: real pricing data for 25+ models (OpenAI, Anthropic,
DeepSeek, Google, Meta) with fuzzy model name matching
- Configurable time window: --days flag (default 30)
- Source filtering: --source flag to filter by platform
Three entry points:
- /insights slash command in CLI (supports --days and --source flags)
- /insights slash command in gateway (compact markdown format)
- hermes insights CLI subcommand (standalone)
Includes 56 tests covering pricing helpers, format helpers, empty DB,
populated DB with multi-platform data, filtering, formatting, and edge cases.
Authored by Farukest. Fixes#432. Extracts _kill_port_process() helper
that uses netstat+taskkill on Windows and fuser on Linux. Previously,
fuser calls were inline with bare except-pass, so on Windows orphaned
bridge processes were never cleaned up — causing 'address already in use'
errors on reconnect. Includes 5 tests covering both platforms, port
matching edge cases, and exception suppression.
Authored by Farukest. Fixes#435. The retry summary in
_handle_max_iterations() hardcoded max_tokens instead of using
_max_tokens_param(), which returns max_completion_tokens for direct
OpenAI API (required by gpt-4o, o-series). The first attempt already
used _max_tokens_param correctly — only the retry path was wrong.
Includes 4 tests for _max_tokens_param provider detection.
Verifies explicit allowlist keys, catch-all _API_KEY/_TOKEN patterns,
case insensitivity, TERMINAL_SSH prefix, and config.yaml routing for
non-secret keys. Covers the fix from PR #469.
The mock handler checked for function_name == 'search' but the RPC
sends 'search_files'. Any test exercising search_files through the
mock would get 'Unknown tool' instead of the canned response.
The _TOOL_STUBS dict in code_execution_tool.py was out of sync with the
actual tool schemas, causing TypeErrors when the LLM used parameters it
sees in its system prompt but the sandbox stubs didn't accept:
search_files:
- Added missing params: context, offset, output_mode
- Fixed target default: 'grep' → 'content' (old value was obsolete)
patch:
- Added missing params: mode, patch (V4A multi-file patch support)
Also added 4 drift-detection tests (TestStubSchemaDrift) that will
catch future divergence between stubs and real schemas:
- test_stubs_cover_all_schema_params: every schema param in stub
- test_stubs_pass_all_params_to_rpc: every stub param sent over RPC
- test_search_files_target_uses_current_values: no obsolete values
- test_generated_module_accepts_all_params: generated code compiles
All 28 tests pass.
Authored by rovle. Adds Daytona as the sixth terminal execution backend
with cloud sandboxes, persistent workspaces, and full CLI/gateway integration.
Includes 24 unit tests and 8 integration tests.
The execute_code sandbox generates a hermes_tools.py stub module for LLM
scripts. Three common failure modes keep tripping up scripts:
1. json.loads(strict=True) rejects control chars in terminal() output
(e.g., GitHub issue bodies with literal tabs/newlines)
2. Shell backtick/quote interpretation when interpolating dynamic content
into terminal() commands (markdown with backticks gets eaten by bash)
3. No retry logic for transient network failures (API timeouts, rate limits)
Adds three convenience helpers to the generated hermes_tools module:
- json_parse(text) — json.loads with strict=False for tolerant parsing
- shell_quote(s) — shlex.quote() for safe shell interpolation
- retry(fn, max_attempts=3, delay=2) — exponential backoff wrapper
Also updates the EXECUTE_CODE_SCHEMA description to document these helpers
so LLMs know they're available without importing anything extra.
Includes 7 new tests (unit + integration) covering all three helpers.
The original implementation only supported xclip (X11), which silently
fails on WSL2 (can't access Windows clipboard for images), Wayland
desktops (xclip is X11-only), and VSCode terminal on WSL2.
Clipboard backend changes (hermes_cli/clipboard.py):
- WSL2: detect via /proc/version, use powershell.exe with .NET
System.Windows.Forms.Clipboard to extract images as base64 PNG
- Wayland: use wl-paste with MIME type detection, auto-convert BMP
to PNG for WSLg environments (via Pillow or ImageMagick)
- Dispatch order: WSL → Wayland → X11 (xclip), with fallthrough
- New has_clipboard_image() for lightweight clipboard checks
- Cache WSL detection result per-process
CLI changes (cli.py):
- /paste command: explicit clipboard image check for terminals where
BracketedPaste doesn't fire (image-only clipboard in VSCode/WinTerm)
- Ctrl+V keybinding: fallback for Linux terminals where Ctrl+V sends
raw byte instead of triggering bracketed paste
Tests: 80 tests (up from 37) covering WSL, Wayland, X11 dispatch,
BMP conversion, has_clipboard_image, and /paste command.
Copy an image to clipboard (screenshot, browser, etc.) and paste into
the Hermes CLI. The image is saved to ~/.hermes/images/, shown as a
badge above the input ([📎 Image #1]), and sent to the model as a
base64-encoded OpenAI vision multimodal content block.
Implementation:
- hermes_cli/clipboard.py: clean module with platform-specific extraction
- macOS: pngpaste (if installed) → osascript fallback (always available)
- Linux: xclip (apt install xclip)
- cli.py: BracketedPaste key handler checks clipboard on every paste,
image bar widget shows attached images, chat() converts to multimodal
content format, Ctrl+C clears attachments
Inspired by @m0at's fork (https://github.com/m0at/hermes-agent) which
implemented image paste support for local vision models. Reimplemented
cleanly as a separate module with tests.
On top of PR #460: self-hosted Firecrawl instances don't require an API
key (USE_DB_AUTHENTICATION=false), so don't force users to set a dummy
FIRECRAWL_API_KEY when FIRECRAWL_API_URL is set. Also adds a proper
self-hosting section to the configuration docs explaining what you get,
what you lose, and how to set it up (Docker stack, tradeoffs vs cloud).
Added 2 more tests (URL-only without key, neither-set raises).
Replaces the unsafe 128K fallback for unknown models with a descending
probe strategy (2M → 1M → 512K → 200K → 128K → 64K → 32K). When a
context-length error occurs, the agent steps down tiers and retries.
The discovered limit is cached per model+provider combo in
~/.hermes/context_length_cache.yaml so subsequent sessions skip probing.
Also parses API error messages to extract the actual context limit
(e.g. 'maximum context length is 32768 tokens') for instant resolution.
The CLI banner now displays the context window size next to the model
name (e.g. 'claude-opus-4 · 200K context · Nous Research').
Changes:
- agent/model_metadata.py: CONTEXT_PROBE_TIERS, persistent cache
(save/load/get), parse_context_limit_from_error(), get_next_probe_tier()
- agent/context_compressor.py: accepts base_url, passes to metadata
- run_agent.py: step-down logic in context error handler, caches on success
- cli.py + hermes_cli/banner.py: context length in welcome banner
- tests: 22 new tests for probing, parsing, and caching
Addresses #132. PR #319's approach (8K default) rejected — too conservative.
