Add unauthorized_dm_behavior config (pair|ignore) with global default
and per-platform override. WhatsApp can silently drop unknown DMs
instead of sending pairing codes.
Adapted config bridging to work with gw_data dict (pre-construction)
rather than config object. Dropped implementation plan document.
Co-authored-by: Frederico Ribeiro <fr@tecompanytea.com>
* Improve tool batching independence checks
* fix: address review feedback on path-aware batching
- Log malformed/non-dict tool arguments at debug level before
falling back to sequential, instead of silently swallowing
the error into an empty dict
- Guard empty paths in _paths_overlap (unreachable in practice
due to upstream filtering, but makes the invariant explicit)
- Add tests: malformed JSON args, non-dict args, _paths_overlap
unit tests including empty path edge cases
- web_crawl is not a registered tool (only web_search/web_extract
are); no addition needed to _PARALLEL_SAFE_TOOLS
---------
Co-authored-by: kshitij <82637225+kshitijk4poor@users.noreply.github.com>
- Strip '_tools' suffix from internal toolset identifiers in the banner
(e.g. 'web_tools' -> 'web', 'homeassistant_tools' -> 'homeassistant')
- Stop appending '_tools' to unavailable toolset names
- Replace 6 hardcoded hex colors (#B8860B, #FFBF00, #FFF8DC) in toolset
rows, overflow line, and MCP server rows with the skin variables
(dim, accent, text) already resolved at the top of the function
Inspired by PR #1871 by @kshitijk4poor.
Adds 4 tests.
* fix: banner skill count now respects disabled skills and platform filtering
The banner's get_available_skills() was doing a raw rglob scan of
~/.hermes/skills/ without checking:
- Whether skills are disabled (skills.disabled config)
- Whether skills match the current platform (platforms: frontmatter)
This caused the banner to show inflated skill counts (e.g. '100 skills'
when many are disabled) and list macOS-only skills on Linux.
Fix: delegate to _find_all_skills() from tools/skills_tool which already
handles both platform gating and disabled-skill filtering.
* fix: system prompt and slash commands now respect disabled skills
Two more places where disabled skills were still surfaced:
1. build_skills_system_prompt() in prompt_builder.py — disabled skills
appeared in the <available_skills> system prompt section, causing
the agent to suggest/load them despite being disabled.
2. scan_skill_commands() in skill_commands.py — disabled skills still
registered as /skill-name slash commands in CLI help and could be
invoked.
Both now load _get_disabled_skill_names() and filter accordingly.
* fix: skill_view blocks disabled skills
skill_view() checked platform compatibility but not disabled state,
so the agent could still load and read disabled skills directly.
Now returns a clear error when a disabled skill is requested, telling
the user to enable it via hermes skills or inspect the files manually.
---------
Co-authored-by: Test <test@test.com>
Recognize hermes_cli/main.py gateway command lines in gateway
process detection and PID validation so --replace reliably finds
existing gateway instances.
Adds a regression test covering script-style cmdline detection.
Closes#1830
Each configured MCP server now registers as its own toolset in TOOLSETS
(e.g. TOOLSETS['github'] = {tools: ['mcp_github_list_files', ...]}),
making raw server names resolvable in platform_toolsets overrides.
Previously MCP tools were only injected into hermes-* umbrella toolsets,
so gateway sessions using raw toolset names like ['terminal', 'github']
in platform_toolsets couldn't resolve MCP tools.
Skips server names that collide with built-in toolsets. Also handles
idempotent reloads (syncs toolsets even when no new servers connect).
Inspired by PR #1876 by @kshitijk4poor.
Adds 2 tests (standalone toolset creation + built-in collision guard).
* perf: cache base_url.lower() via property, consolidate triple load_config(), hoist set constant
run_agent.py:
- Add base_url property that auto-caches _base_url_lower on every
assignment, eliminating 12+ redundant .lower() calls per API cycle
across __init__, _build_api_kwargs, _supports_reasoning_extra_body,
and the main conversation loop
- Consolidate three separate load_config() disk reads in __init__
(memory, skills, compression) into a single call, reusing the
result dict for all three config sections
model_tools.py:
- Hoist _READ_SEARCH_TOOLS set to module level (was rebuilt inside
handle_function_call on every tool invocation)
* Use endpoint metadata for custom model context and pricing
---------
Co-authored-by: kshitij <82637225+kshitijk4poor@users.noreply.github.com>
Add _wait_for_gateway_exit() that polls get_running_pid() to confirm
the old gateway process has actually exited before starting a new one.
If the process doesn't exit within 5s, sends SIGKILL to the specific
PID. Uses the saved PID from gateway.pid (not launchd labels) so it
works correctly with multiple gateway instances under separate
HERMES_HOME directories.
Applied to both launchd_restart() and the manual restart path (replaces
the blind time.sleep(2)).
Inspired by PR #1881 by @AzothZephyr (race condition diagnosis).
Adds 4 tests.
When config.yaml had a non-default model (e.g. gpt-5.3-codex) and the
provider was openai-codex, _normalize_model_for_provider() would replace
it with the latest available codex model because _model_is_default only
checked the CLI argument, not the config value.
Now _model_is_default is False when config.yaml has a model that differs
from the global fallback (anthropic/claude-opus-4.6), so the user's
explicit config choice is preserved.
Fixes#1887
Co-authored-by: Test <test@test.com>
MiniMax: Add M2.7 and M2.7-highspeed as new defaults across provider
model lists, auxiliary client, metadata, setup wizard, RL training tool,
fallback tests, and docs. Retain M2.5/M2.1 as alternatives.
OpenRouter: Add grok-4.20-beta, nemotron-3-super-120b-a12b:free,
trinity-large-preview:free, glm-5-turbo, and hunter-alpha to the
model catalog.
MiniMax changes based on PR #1882 by @octo-patch (applied manually
due to stale conflicts in refactored pricing module).
* fix: include ACP sessions in default search sources
* fix: remove hardcoded source allowlist from session search
The default source_filter was a hardcoded list that silently excluded
any platform not explicitly listed. Instead of maintaining an ever-growing
allowlist, remove it entirely so all sources are searched by default.
Callers can still pass source_filter explicitly to narrow results.
Follow-up to cherry-picked PR #1817.
---------
Co-authored-by: someoneexistsontheinternet <154079416+someoneexistsontheinternet@users.noreply.github.com>
Co-authored-by: Test <test@test.com>
- Update _is_anthropic_oauth in _try_refresh_anthropic_client_credentials()
when token type changes during credential refresh
- Set _is_anthropic_oauth in _try_activate_fallback() Anthropic path
- Move _turns_since_memory and _iters_since_skill init to __init__ so
nudge counters accumulate across run_conversation() calls in CLI mode
- Remove unreachable retry_count >= max_retries block after raise
Adds 7 regression tests. Salvaged from PR #1797 by @0xbyt4.
Agent-created skills were using the same policy as community hub
installs, blocking any skill with medium/high severity findings
(e.g. docker pull, pip install, git clone). This meant the agent
couldn't create skills that reference Docker or other common tools.
Changed agent-created policy from (allow, block, block) to
(allow, allow, block) — matching the trusted policy. Caution-level
findings (medium/high severity) are now allowed through, while
dangerous findings (critical severity like exfiltration, prompt
injection, reverse shells) remain blocked.
Added 4 tests covering the agent-created policy: safe allowed,
caution allowed, dangerous blocked, force override.
Every cron job prompt now includes guidance that the agent can respond
with [SILENT] when it has nothing new or noteworthy to report. The
scheduler checks for this marker and skips delivery, while still saving
output to disk for audit. Failed jobs always deliver regardless.
This replaces the notify parameter approach from PR #1807 with a simpler
always-on design — the model is smart enough to decide when there's
nothing worth reporting without needing a per-job flag.
* feat: OpenAI-compatible API server platform adapter
Salvaged from PR #956, updated for current main.
Adds an HTTP API server as a gateway platform adapter that exposes
hermes-agent via the OpenAI Chat Completions and Responses APIs.
Any OpenAI-compatible frontend (Open WebUI, LobeChat, LibreChat,
AnythingLLM, NextChat, ChatBox, etc.) can connect by pointing at
http://localhost:8642/v1.
Endpoints:
- POST /v1/chat/completions — stateless Chat Completions API
- POST /v1/responses — stateful Responses API with chaining
- GET /v1/responses/{id} — retrieve stored response
- DELETE /v1/responses/{id} — delete stored response
- GET /v1/models — list hermes-agent as available model
- GET /health — health check
Features:
- Real SSE streaming via stream_delta_callback (uses main's streaming)
- In-memory LRU response store for Responses API conversation chaining
- Named conversations via 'conversation' parameter
- Bearer token auth (optional, via API_SERVER_KEY)
- CORS support for browser-based frontends
- System prompt layering (frontend system messages on top of core)
- Real token usage tracking in responses
Integration points:
- Platform.API_SERVER in gateway/config.py
- _create_adapter() branch in gateway/run.py
- API_SERVER_* env vars in hermes_cli/config.py
- Env var overrides in gateway/config.py _apply_env_overrides()
Changes vs original PR #956:
- Removed streaming infrastructure (already on main via stream_consumer.py)
- Removed Telegram reply_to_mode (separate feature, not included)
- Updated _resolve_model() -> _resolve_gateway_model()
- Updated stream_callback -> stream_delta_callback
- Updated connect()/disconnect() to use _mark_connected()/_mark_disconnected()
- Adapted to current Platform enum (includes MATTERMOST, MATRIX, DINGTALK)
Tests: 72 new tests, all passing
Docs: API server guide, Open WebUI integration guide, env var reference
* feat(whatsapp): make reply prefix configurable via config.yaml
Reworked from PR #1764 (ifrederico) to use config.yaml instead of .env.
The WhatsApp bridge prepends a header to every outgoing message.
This was hardcoded to '⚕ *Hermes Agent*'. Users can now customize
or disable it via config.yaml:
whatsapp:
reply_prefix: '' # disable header
reply_prefix: '🤖 *My Bot*\n───\n' # custom prefix
How it works:
- load_gateway_config() reads whatsapp.reply_prefix from config.yaml
and stores it in PlatformConfig.extra['reply_prefix']
- WhatsAppAdapter reads it from config.extra at init
- When spawning bridge.js, the adapter passes it as
WHATSAPP_REPLY_PREFIX in the subprocess environment
- bridge.js handles undefined (default), empty (no header),
or custom values with \\n escape support
- Self-chat echo suppression uses the configured prefix
Also fixes _config_version: was 9 but ENV_VARS_BY_VERSION had a
key 10 (TAVILY_API_KEY), so existing users at v9 would never be
prompted for Tavily. Bumped to 10 to close the gap. Added a
regression test to prevent this from happening again.
Credit: ifrederico (PR #1764) for the bridge.js implementation
and the config version gap discovery.
---------
Co-authored-by: Test <test@test.com>
Save and restore the process-global _last_resolved_tool_names in
_run_single_child() so the parent's execute_code sandbox generates
correct tool imports after delegation completes.
The global was already mostly mitigated (run_agent.py passes
enabled_tools via self.valid_tool_names), but the global itself
remained corrupted — a footgun for any code that reads it directly.
Co-authored-by: shane9coy <shane9coy@users.noreply.github.com>
* fix(session): skip corrupt lines in load_transcript instead of crashing
Wrap json.loads() in load_transcript() with try/except JSONDecodeError
so that partial JSONL lines (from mid-write crashes like OOM/SIGKILL)
are skipped with a warning instead of crashing the entire transcript
load. The rest of the history loads fine.
Adds a logger.warning with the session ID and truncated corrupt line
content for debugging visibility.
Salvaged from PR #1193 by alireza78a.
Closes#1193
* fix(stt): respect explicit provider config instead of env-var fallback
Rework _get_provider() to separate explicit config from auto-detect.
When stt.provider is explicitly set in config.yaml, that choice is
authoritative — no silent cross-provider fallback based on which env
vars happen to be set. When no provider is configured, auto-detect
still tries: local > groq > openai.
This fixes the reported scenario where provider: local + a placeholder
OPENAI_API_KEY caused the system to silently select OpenAI and fail
with a 401.
Closes#1774
_sanitize_fts5_query() was stripping ALL double quotes (including
properly paired ones), breaking user-provided quoted phrases like
"exact phrase". Hyphenated terms like chat-send also silently
expanded to chat AND send, returning unexpected or zero results.
Fix:
1. Extract balanced quoted phrases into placeholders before
stripping FTS5-special characters, then restore them.
2. Wrap unquoted hyphenated terms (word-word) in double quotes so
FTS5 matches them as exact phrases instead of splitting on
the hyphen.
3. Unmatched quotes are still stripped as before.
Based on issue report by @bailob (#1770) and PR #1773 by @Jah-yee
(whose branch contained unrelated changes and couldn't be merged
directly).
Closes#1770Closes#1773
Co-authored-by: Jah-yee <Jah-yee@users.noreply.github.com>
compress() checks both the head and tail neighbors when choosing the
summary message role. When only the tail collides, the role is flipped.
When BOTH roles would create consecutive same-role messages (e.g.
head=assistant, tail=user), the summary is merged into the first tail
message instead of inserting a standalone message that breaks role
alternation and causes API 400 errors.
The previous code handled head-side collision but left the tail-side
uncovered — long conversations would crash mid-reply with no useful
error, forcing the user to /reset and lose session history.
Based on PR #1186 by @alireza78a, with improved double-collision
handling (merge into tail instead of unconditional 'user' fallback).
Co-authored-by: alireza78a <alireza78.crypto@gmail.com>
Wrap json.loads() in load_transcript() with try/except JSONDecodeError
so that partial JSONL lines (from mid-write crashes like OOM/SIGKILL)
are skipped with a warning instead of crashing the entire transcript
load. The rest of the history loads fine.
Adds a logger.warning with the session ID and truncated corrupt line
content for debugging visibility.
Salvaged from PR #1193 by alireza78a.
Closes#1193
- Add summary_base_url config option to compression block for custom
OpenAI-compatible endpoints (e.g. zai, DeepSeek, Ollama)
- Remove compression env var bridges from cli.py and gateway/run.py
(CONTEXT_COMPRESSION_* env vars no longer set from config)
- Switch run_agent.py to read compression config directly from
config.yaml instead of env vars
- Fix backwards-compat block in _resolve_task_provider_model to also
fire when auxiliary.compression.provider is 'auto' (DEFAULT_CONFIG
sets this, which was silently preventing the compression section's
summary_* keys from being read)
- Add test for summary_base_url config-to-client flow
- Update docs to show compression as config.yaml-only
Closes#1591
Based on PR #1702 by @uzaylisak
Salvage of PR #1707 by @kshitijk4poor (cherry-picked with authorship preserved).
Adds Tavily as a third web backend alongside Firecrawl and Parallel, using the Tavily REST API via httpx.
- Backend selection via hermes tools → saved as web.backend in config.yaml
- All three tools supported: search, extract, crawl
- TAVILY_API_KEY in config registry, doctor, status, setup wizard
- 15 new Tavily tests + 9 backend selection tests + 5 config tests
- Backward compatible
Closes#1707
Salvage of PR #1321 by @alireza78a (cherry-picked concept, reimplemented
against current main).
Phase 1 — Pre-call message sanitization:
_sanitize_api_messages() now runs unconditionally before every LLM call.
Previously gated on context_compressor being present, so sessions loaded
from disk or running without compression could accumulate dangling
tool_call/tool_result pairs causing API errors.
Phase 2a — Delegate task cap:
_cap_delegate_task_calls() truncates excess delegate_task calls per turn
to MAX_CONCURRENT_CHILDREN. The existing cap in delegate_tool.py only
limits the task array within a single call; this catches multiple
separate delegate_task tool_calls in one turn.
Phase 2b — Tool call deduplication:
_deduplicate_tool_calls() drops duplicate (tool_name, arguments) pairs
within a single turn when models stutter.
All three are static methods on AIAgent, independently testable.
29 tests covering happy paths and edge cases.
Adds .hermes.md / HERMES.md discovery for per-project agent configuration.
When the agent starts, it walks from cwd to the git root looking for
.hermes.md (preferred) or HERMES.md, strips any YAML frontmatter, and
injects the markdown body into the system prompt as project context.
- Nearest-first discovery (subdirectory configs shadow parent)
- Stops at git root boundary (no leaking into parent repos)
- YAML frontmatter stripped (structured config deferred to Phase 2)
- Same injection scanning and 20K truncation as other context files
- 22 comprehensive tests
Original implementation by ch3ronsa. Cherry-picked and adapted for current main.
Closes#681 (Phase 1)
After the first user→assistant exchange, Hermes now generates a short
descriptive session title via the auxiliary LLM (compression task config).
Title generation runs in a background thread so it never delays the
user-facing response.
Key behaviors:
- Fires only on the first 1-2 exchanges (checks user message count)
- Skips if a title already exists (user-set titles are never overwritten)
- Uses call_llm with compression task config (cheapest/fastest model)
- Truncates long messages to keep the title generation request small
- Cleans up LLM output: strips quotes, 'Title:' prefixes, enforces 80 char max
- Works in both CLI and gateway (Telegram/Discord/etc.)
Also updates /title (no args) to show the session ID alongside the title
in both CLI and gateway.
Implements #1426
* feat(web): add Parallel as alternative web search/extract backend
Adds Parallel (parallel.ai) as a drop-in alternative to Firecrawl for
web_search and web_extract tools using the official parallel-web SDK.
- Backend selection via WEB_SEARCH_BACKEND env var (auto/parallel/firecrawl)
- Auto mode prefers Firecrawl when both keys present; Parallel when sole backend
- web_crawl remains Firecrawl-only with clear error when unavailable
- Lazy SDK imports, interrupt support, singleton clients
- 16 new unit tests for backend selection and client config
Co-authored-by: s-jag <s-jag@users.noreply.github.com>
* fix: add PARALLEL_API_KEY to config registry and fix web_crawl policy tests
Follow-up for Parallel backend integration:
- Add PARALLEL_API_KEY to OPTIONAL_ENV_VARS (hermes doctor, env blocklist)
- Add to set_config_value api_keys list (hermes config set)
- Add to doctor keys display
- Fix 2 web_crawl policy tests that didn't set FIRECRAWL_API_KEY
(needed now that web_crawl has a Firecrawl availability guard)
* refactor: explicit backend selection via hermes tools, not auto-detect
Replace the auto-detect backend selection with explicit user choice:
- hermes tools saves WEB_SEARCH_BACKEND to .env when user picks a provider
- _get_backend() reads the explicit choice first
- Fallback only for manual/legacy config (uses whichever key is present)
- _is_provider_active() shows [active] for the selected web backend
- Updated tests, docs, and .env.example to remove 'auto' mode language
* refactor: use config.yaml for web backend, not env var
Match the TTS/browser pattern — web.backend is stored in config.yaml
(set by hermes tools), not as a WEB_SEARCH_BACKEND env var.
- _load_web_config() reads web: section from config.yaml
- _get_backend() reads web.backend from config, falls back to key detection
- _configure_provider() saves to config dict (saved to config.yaml)
- _is_provider_active() reads from config dict
- Removed WEB_SEARCH_BACKEND from .env.example, set_config_value, docs
- Updated all tests to mock _load_web_config instead of env vars
---------
Co-authored-by: s-jag <s-jag@users.noreply.github.com>
When container_persistent=false, the inner mini-swe-agent cleanup only
runs 'docker stop' in the background, leaving containers in Exited state.
Now cleanup() also runs 'docker rm -f' to fully remove the container.
Also fixes pre-existing test failures in model_metadata (gpt-4.1 1M context),
setup tests (TTS provider step), and adds MockInnerDocker.cleanup().
Original fix by crazywriter1. Cherry-picked and adapted for current main.
Fixes#1679
Salvaged from PR #1708 by @kartikkabadi. Cherry-picked with authorship preserved.
Fixes pre-existing test failures from setup TTS prompt flow changes and environment-sensitive assumptions.
Co-authored-by: Kartik <user2@RentKars-MacBook-Air.local>
* feat: interactive MCP tool configuration in hermes tools
Add the ability to selectively enable/disable individual MCP server
tools through the interactive 'hermes tools' TUI.
Changes:
- tools/mcp_tool.py: Add probe_mcp_server_tools() — lightweight function
that temporarily connects to configured MCP servers, discovers their
tools (names + descriptions), and disconnects. No registry side effects.
- hermes_cli/tools_config.py: Add 'Configure MCP tools' option to the
interactive menu. When selected:
1. Probes all enabled MCP servers for their available tools
2. Shows a per-server curses checklist with tool descriptions
3. Pre-selects tools based on existing include/exclude config
4. Writes changes back as tools.exclude entries in config.yaml
5. Reports which servers failed to connect
The existing CLI commands (hermes tools enable/disable server:tool)
continue to work unchanged. This adds the interactive TUI counterpart
so users can browse and toggle MCP tools visually.
Tests: 22 new tests covering probe function edge cases and interactive
flow (pre-selection, exclude/include modes, description truncation,
multi-server handling, error paths).
* feat(telegram): auto-detect HTML tags and use parse_mode=HTML in send_message
When _send_telegram detects HTML tags in the message body, it now sends
with parse_mode='HTML' instead of converting to MarkdownV2. This allows
cron jobs and agents to send rich HTML-formatted Telegram messages with
bold, italic, code blocks, etc. that render correctly.
Detection uses the same regex from PR #1568 by @ashaney:
re.search(r'<[a-zA-Z/][^>]*>', message)
Plain-text and markdown messages continue through the existing
MarkdownV2 pipeline. The HTML fallback path also catches HTML parse
errors and falls back to plain text, matching the existing MarkdownV2
error handling.
Inspired by: github.com/ashaney — PR #1568
Salvaged from PR #1573 by @eren-karakus0. Cherry-picked with authorship preserved.
Fixes#1143 — background process notifications resume after gateway restart.
Co-authored-by: Muhammet Eren Karakuş <erenkar950@gmail.com>
Add the ability to selectively enable/disable individual MCP server
tools through the interactive 'hermes tools' TUI.
Changes:
- tools/mcp_tool.py: Add probe_mcp_server_tools() — lightweight function
that temporarily connects to configured MCP servers, discovers their
tools (names + descriptions), and disconnects. No registry side effects.
- hermes_cli/tools_config.py: Add 'Configure MCP tools' option to the
interactive menu. When selected:
1. Probes all enabled MCP servers for their available tools
2. Shows a per-server curses checklist with tool descriptions
3. Pre-selects tools based on existing include/exclude config
4. Writes changes back as tools.exclude entries in config.yaml
5. Reports which servers failed to connect
The existing CLI commands (hermes tools enable/disable server:tool)
continue to work unchanged. This adds the interactive TUI counterpart
so users can browse and toggle MCP tools visually.
Tests: 22 new tests covering probe function edge cases and interactive
flow (pre-selection, exclude/include modes, description truncation,
multi-server handling, error paths).
- Default enabled: false (zero overhead when not configured)
- Fast path: cached disabled state skips all work immediately
- TTL cache (30s) for parsed policy — avoids re-reading config.yaml
on every URL check
- Missing shared files warn + skip instead of crashing all web tools
- Lazy yaml import — missing PyYAML doesn't break browser toolset
- Guarded browser_tool import — fail-open lambda fallback
- check_website_access never raises for default path (fail-open with
warning log); only raises with explicit config_path (test mode)
- Simplified enforcement code in web_tools/browser_tool — no more
try/except wrappers since errors are handled internally
Add DingTalk as a messaging platform using the dingtalk-stream SDK
for real-time message reception via Stream Mode (no webhook needed).
Replies are sent via session webhook using markdown format.
Features:
- Stream Mode connection (long-lived WebSocket, no public URL needed)
- Text and rich text message support
- DM and group chat support
- Message deduplication with 5-minute window
- Auto-reconnection with exponential backoff
- Session webhook caching for reply routing
Configuration:
export DINGTALK_CLIENT_ID=your-app-key
export DINGTALK_CLIENT_SECRET=your-app-secret
# or in config.yaml:
platforms:
dingtalk:
enabled: true
extra:
client_id: your-app-key
client_secret: your-app-secret
Files:
- gateway/platforms/dingtalk.py (340 lines) — adapter implementation
- gateway/config.py — add DINGTALK to Platform enum
- gateway/run.py — add DingTalk to _create_adapter
- hermes_cli/config.py — add env vars to _EXTRA_ENV_KEYS
- hermes_cli/tools_config.py — add dingtalk to PLATFORMS
- tests/gateway/test_dingtalk.py — 21 tests
