fix(gateway): preserve scoped Telegram authorization - #84141
Closed
miroslavb wants to merge 3 commits into
Closed
Conversation
Restore captured multiplex profile scope inside adapter authorization callbacks, route Telegram intake and approval checks through the registered gateway callback, and ignore synthetic systemd manager entries unless LoadState is loaded.
alt-glitch
added
type/bug
Collaborator
bob798
pushed a commit
to bob798/hermes-agent
that referenced
this pull request
Aug 12, 2026
Adds Tool Search, a structured-tools progressive-disclosure layer that
replaces MCP and non-core plugin tools in the model-visible tools array
with three bridge tools (tool_search / tool_describe / tool_call) when
the deferrable surface would consume more than a configurable percentage
of the active model's context window. Core Hermes tools are never deferred.
Default mode is 'auto' with a 10% context threshold, so small toolsets
pay no overhead. Set tools.tool_search.enabled to 'on' to force or 'off'
to disable.
Design carefully reflects the OpenClaw production failure modes
documented in the openclaw-tool-search-report:
- Core tools never defer (toolsets._HERMES_CORE_TOOLS). Addresses the
'tools silently missing from isolated cron turns' regression class
(openclaw#84141) by construction: there is no code path that can
drop a core tool.
- Catalog is stateless across turns — rebuilt from the live tool-defs
list on every assembly. No session-keyed Map that can drift out of
sync with the registry.
- tool_call unwraps the bridge call before any hook fires, so plugin
pre/post hooks, guardrails, approval flows, and the activity feed
all see the underlying tool name, not the bridge (addresses
openclaw#85588 and the verbose-mode complaint on openclaw#79823).
- The unwrap happens in both the parallel and sequential paths of
agent/tool_executor.py and also in handle_function_call, so direct
callers (sandboxed code, eval harnesses) are covered too.
- Bridge tools cannot invoke each other (recursion guard) and cannot
invoke core tools (those must be called directly).
- Tools mode only — no JS-sandbox code-mode. Keeps the surface small.
- Token estimation via cheap char/4 heuristic; precision isn't needed
for the threshold decision.
Files:
- tools/tool_search.py — new module (BM25 retrieval, classification,
threshold gate, bridge dispatch, unwrap helper).
- tests/tools/test_tool_search.py — 35 tests including the OpenClaw
NousResearch#84141 regression guard.
- model_tools.py — wires assembly into _compute_tool_definitions as the
final step, adds skip_tool_search_assembly kwarg so the bridge can
see the real catalog, dispatches the three bridge tools.
- agent/tool_executor.py — unwraps tool_call in both parallel and
sequential parsing loops so checkpointing, guardrails, plugin hooks,
and tool-progress callbacks all observe the underlying tool name.
- hermes_cli/config.py — DEFAULT_CONFIG['tools']['tool_search'] block.
- website/docs/user-guide/features/tool-search.md — user docs.
Validation:
- 35/35 new tests pass.
- Existing tool/registry/model_tools/config/coercion/executor tests
(82 + 74 + small adjacents) green.
- Live E2E: 20 fake MCP tools registered, get_tool_definitions returns
3 bridges, tool_search returns top 3 hits, tool_describe returns
full schema, tool_call dispatches to the real underlying handler
and the underlying result is what the model sees.
- Reserved-name recursion guard verified live.
- Core-tool refusal via tool_call verified live.
added 2 commits
August 14, 2026 06:47
Resolve stale compression roots through one coherent SQLite snapshot, revalidate the live tip atomically with transcript insertion, and commit in-memory adoption only after durable writes succeed.
19 tasks
Author
|
Closing this PR rather than asking maintainers to review two unrelated changes on one conflicting branch.\n\nThe multi-generation compression/session-persistence fix has been extracted onto current main as #86099, which contains one scoped commit and formally Fixes #82001.\n\nThe original Telegram authorization change from commit 9a5ab2b is not part of #86099 and is not being represented as completed there. If that authorization fix is still needed after reconciling the overlap noted with #65589 and #34734, it should be rebased and submitted as its own dedicated PR.\n\nNo reviews or reported checks are being discarded here; this PR had none. |
skappafrost
pushed a commit
to skappafrost/hermes-agent
that referenced
this pull request
Aug 15, 2026
Adds Tool Search, a structured-tools progressive-disclosure layer that
replaces MCP and non-core plugin tools in the model-visible tools array
with three bridge tools (tool_search / tool_describe / tool_call) when
the deferrable surface would consume more than a configurable percentage
of the active model's context window. Core Hermes tools are never deferred.
Default mode is 'auto' with a 10% context threshold, so small toolsets
pay no overhead. Set tools.tool_search.enabled to 'on' to force or 'off'
to disable.
Design carefully reflects the OpenClaw production failure modes
documented in the openclaw-tool-search-report:
- Core tools never defer (toolsets._HERMES_CORE_TOOLS). Addresses the
'tools silently missing from isolated cron turns' regression class
(openclaw#84141) by construction: there is no code path that can
drop a core tool.
- Catalog is stateless across turns — rebuilt from the live tool-defs
list on every assembly. No session-keyed Map that can drift out of
sync with the registry.
- tool_call unwraps the bridge call before any hook fires, so plugin
pre/post hooks, guardrails, approval flows, and the activity feed
all see the underlying tool name, not the bridge (addresses
openclaw#85588 and the verbose-mode complaint on openclaw#79823).
- The unwrap happens in both the parallel and sequential paths of
agent/tool_executor.py and also in handle_function_call, so direct
callers (sandboxed code, eval harnesses) are covered too.
- Bridge tools cannot invoke each other (recursion guard) and cannot
invoke core tools (those must be called directly).
- Tools mode only — no JS-sandbox code-mode. Keeps the surface small.
- Token estimation via cheap char/4 heuristic; precision isn't needed
for the threshold decision.
Files:
- tools/tool_search.py — new module (BM25 retrieval, classification,
threshold gate, bridge dispatch, unwrap helper).
- tests/tools/test_tool_search.py — 35 tests including the OpenClaw
NousResearch#84141 regression guard.
- model_tools.py — wires assembly into _compute_tool_definitions as the
final step, adds skip_tool_search_assembly kwarg so the bridge can
see the real catalog, dispatches the three bridge tools.
- agent/tool_executor.py — unwraps tool_call in both parallel and
sequential parsing loops so checkpointing, guardrails, plugin hooks,
and tool-progress callbacks all observe the underlying tool name.
- hermes_cli/config.py — DEFAULT_CONFIG['tools']['tool_search'] block.
- website/docs/user-guide/features/tool-search.md — user docs.
Validation:
- 35/35 new tests pass.
- Existing tool/registry/model_tools/config/coercion/executor tests
(82 + 74 + small adjacents) green.
- Live E2E: 20 fake MCP tools registered, get_tool_definitions returns
3 bridges, tool_search returns top 3 hits, tool_describe returns
full schema, tool_call dispatches to the real underlying handler
and the underlying result is what the model sees.
- Reserved-name recursion guard verified live.
- Core-tool refusal via tool_call verified live.
prmartinow
pushed a commit
to prmartinow/hermes-agent
that referenced
this pull request
Aug 26, 2026
Adds Tool Search, a structured-tools progressive-disclosure layer that
replaces MCP and non-core plugin tools in the model-visible tools array
with three bridge tools (tool_search / tool_describe / tool_call) when
the deferrable surface would consume more than a configurable percentage
of the active model's context window. Core Hermes tools are never deferred.
Default mode is 'auto' with a 10% context threshold, so small toolsets
pay no overhead. Set tools.tool_search.enabled to 'on' to force or 'off'
to disable.
Design carefully reflects the OpenClaw production failure modes
documented in the openclaw-tool-search-report:
- Core tools never defer (toolsets._HERMES_CORE_TOOLS). Addresses the
'tools silently missing from isolated cron turns' regression class
(openclaw#84141) by construction: there is no code path that can
drop a core tool.
- Catalog is stateless across turns — rebuilt from the live tool-defs
list on every assembly. No session-keyed Map that can drift out of
sync with the registry.
- tool_call unwraps the bridge call before any hook fires, so plugin
pre/post hooks, guardrails, approval flows, and the activity feed
all see the underlying tool name, not the bridge (addresses
openclaw#85588 and the verbose-mode complaint on openclaw#79823).
- The unwrap happens in both the parallel and sequential paths of
agent/tool_executor.py and also in handle_function_call, so direct
callers (sandboxed code, eval harnesses) are covered too.
- Bridge tools cannot invoke each other (recursion guard) and cannot
invoke core tools (those must be called directly).
- Tools mode only — no JS-sandbox code-mode. Keeps the surface small.
- Token estimation via cheap char/4 heuristic; precision isn't needed
for the threshold decision.
Files:
- tools/tool_search.py — new module (BM25 retrieval, classification,
threshold gate, bridge dispatch, unwrap helper).
- tests/tools/test_tool_search.py — 35 tests including the OpenClaw
NousResearch#84141 regression guard.
- model_tools.py — wires assembly into _compute_tool_definitions as the
final step, adds skip_tool_search_assembly kwarg so the bridge can
see the real catalog, dispatches the three bridge tools.
- agent/tool_executor.py — unwraps tool_call in both parallel and
sequential parsing loops so checkpointing, guardrails, plugin hooks,
and tool-progress callbacks all observe the underlying tool name.
- hermes_cli/config.py — DEFAULT_CONFIG['tools']['tool_search'] block.
- website/docs/user-guide/features/tool-search.md — user docs.
Validation:
- 35/35 new tests pass.
- Existing tool/registry/model_tools/config/coercion/executor tests
(82 + 74 + small adjacents) green.
- Live E2E: 20 fake MCP tools registered, get_tool_definitions returns
3 bridges, tool_search returns top 3 hits, tool_describe returns
full schema, tool_call dispatches to the real underlying handler
and the underlying result is what the model sees.
- Reserved-name recursion guard verified live.
- Core-tool refusal via tool_call verified live.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
systemctl showmanager results unlessLoadState=loadedAGENTS.mdVerification
47 passed— focused systemd/multiplex/Telegram auth/reconnect suite72 passed— multiplex suite523 passed— broad Telegram suite; 6 aggregate order-pollution failures pass in an isolated real-PTB process5224 passed, 29 skipped, 2 xfailed— fulltests/gateway; 8 unrelated failures classified: 3 reproduce on cleanorigin/main, 5 pass in isolated file processesgit diff --check, and static security scan passedRegression contracts
LoadState=not-founduser-manager object cannot shadow the loaded system unit