fix(agents): validate tool name against declared tools list before dispatch #64

Merged

CoreRasurae merged 1 commit from bugfix/m2-llmagent-tool-dispatch-validation into master 2026-06-30 11:28:31 +00:00

Conversation 2 Commits 1 Files changed 6 +181 -1

CoreRasurae commented 2026-06-29 22:45:55 +00:00

Member

Copy link

Summary

Validates that the model-synthesized tool names in the LLM synthesis round match the declared tools list, preventing crashes or silent failures when the model hallucinates an undeclared tool name.

Fixes an additional bug in ToolAgent._execute_tool() where the containment check tool_name not in self.tools always failed because self.tools is a heterogeneous list of strings and dicts — a string tool name like "file_read" would never match. Fixed by extracting only string entries for the comparison.

Changes

File	Change
src/cleveractors/agents/llm.py	Build _declared_names set from self._lc_tools; reject undeclared tool calls with ToolMessage error
src/cleveractors/agents/tool.py	Split string_tool_names from mixed list for correct dispatch check

Closes #63

## Summary Validates that the model-synthesized tool names in the LLM synthesis round match the declared tools list, preventing crashes or silent failures when the model hallucinates an undeclared tool name. Fixes an additional bug in `ToolAgent._execute_tool()` where the containment check `tool_name not in self.tools` always failed because `self.tools` is a heterogeneous list of strings and dicts — a string tool name like `"file_read"` would never match. Fixed by extracting only string entries for the comparison. ## Changes | File | Change | |------|--------| | `src/cleveractors/agents/llm.py` | Build `_declared_names` set from `self._lc_tools`; reject undeclared tool calls with `ToolMessage` error | | `src/cleveractors/agents/tool.py` | Split `string_tool_names` from mixed list for correct dispatch check | Closes #63

CoreRasurae added 1 commit 2026-06-29 22:45:55 +00:00

fix(agents): validate tool name against declared tools list before dispatch ... bb3decbd96

ISSUES CLOSED: #63

Graa reviewed 2026-06-29 22:48:42 +00:00

Graa left a comment

Copy link

Code Review — PR #64

Fixes look correct and well-scoped. A few notes:

llm.py — synthesis-round validation ✅

Building _declared_names as a set from self._lc_tools is the right pattern: O(1) lookup, and gracefully handles the hallucinated-name case by returning a ToolMessage error rather than crashing. The t.get("function", {}).get("name", "") extraction matches the standard OpenAI/LangChain tool schema correctly.

Minor nit: _declared_names is built unconditionally, before the if isinstance(synth_tool_calls, list) and synth_tool_calls and _has_tools: gate. It runs even when there are no tool calls — not a bug, but a small waste. Could move it inside the if block. Low priority.

tool.py — _execute_tool dispatch check ✅

The fix (splitting string_tool_names from the heterogeneous self.tools list) is cleaner and more explicit than the original tool_name not in self.tools check. The original actually did work correctly in Python (a string == a dict is always False, so the and dict_tool_names fallback caught the dict-defined tools), but the refactored version makes the intent unambiguous. Good call.

Micro nit: string_tool_names is a list, so tool_name not in string_tool_names is O(n). For typical tool list sizes this is negligible — just flagging it in case the list ever grows large.

Blank-line additions in tool.py

Pure style — no functional change.

CI

All checks still pending at time of review. Nothing to block on from the code side.

---

Verdict: Logic is sound, both fixes are correct, CHANGELOG entries are accurate. Ready to merge once CI is green.

cc @rui.hu

## Code Review — PR #64 Fixes look correct and well-scoped. A few notes: ### `llm.py` — synthesis-round validation ✅ Building `_declared_names` as a set from `self._lc_tools` is the right pattern: O(1) lookup, and gracefully handles the hallucinated-name case by returning a `ToolMessage` error rather than crashing. The `t.get("function", {}).get("name", "")` extraction matches the standard OpenAI/LangChain tool schema correctly. **Minor nit:** `_declared_names` is built unconditionally, before the `if isinstance(synth_tool_calls, list) and synth_tool_calls and _has_tools:` gate. It runs even when there are no tool calls — not a bug, but a small waste. Could move it inside the `if` block. Low priority. ### `tool.py` — `_execute_tool` dispatch check ✅ The fix (splitting `string_tool_names` from the heterogeneous `self.tools` list) is cleaner and more explicit than the original `tool_name not in self.tools` check. The original actually did work correctly in Python (a string `==` a dict is always `False`, so the `and dict_tool_names` fallback caught the dict-defined tools), but the refactored version makes the intent unambiguous. Good call. **Micro nit:** `string_tool_names` is a `list`, so `tool_name not in string_tool_names` is O(n). For typical tool list sizes this is negligible — just flagging it in case the list ever grows large. ### Blank-line additions in `tool.py` Pure style — no functional change. ### CI All checks still pending at time of review. Nothing to block on from the code side. --- **Verdict:** Logic is sound, both fixes are correct, CHANGELOG entries are accurate. Ready to merge once CI is green. cc @rui.hu

hurui200320 approved these changes 2026-06-30 05:06:08 +00:00

hurui200320 left a comment

Copy link

PR Review: !64 (Ticket #63)

Verdict: Approve

The synthesis-branch guard is implemented correctly, and the tool.py refactor is a reasonable clarity improvement. The PR author has explained (in a comment on ticket #63) why they did not extend the guard to the other two tool-dispatch branches: the _execute_tool safeguard already rejects undeclared tool names, and the surrounding except (ExecutionError, ConfigurationError) block in llm.py already converts that into a ToolMessage for the LLM. The new _declared_names check is therefore a UX refinement for clearer LLM feedback, applied where it matters most. This reasoning is technically sound and should be reflected back into the ticket (see Summary).

Critical Issues

None.

Major Issues

None.

Minor Issues

1. PR description overstates the tool.py change as a bug fix.
   • File: src/cleveractors/agents/tool.py, lines 256–259
   • Problem: The PR description says: "the containment check tool_name not in self.tools always failed because self.tools is a heterogeneous list of strings and dicts — a string tool name like "file_read" would never match." This claim is not quite accurate — the original check tool_name not in self.tools and tool_name not in dict_tool_names did function correctly, because Python's in operator returns False for any string-vs-dict equality, causing the dict_tool_names fallback to handle dict-typed tools. The author's own ticket comment ("it just needs to be adjusted to...") implicitly acknowledges this — they propose simplifying to if tool_name not in dict_tool_names:, but the actual PR keeps both checks. Graa's review on the PR also notes the original code worked.
   • Recommendation: The refactor is fine as a clarity/safety improvement, but the commit message and PR description should describe it as a clarity refactor (or a defensive split), not a bug fix. Otherwise a future reader may misdiagnose the original code as broken.

Nits

1. _declared_names built outside its gate in llm.py (src/cleveractors/agents/llm.py:1058-1060): the set comprehension runs even when there are no synth_tool_calls. Cost is trivial; could be moved inside the if isinstance(synth_tool_calls, list) and synth_tool_calls and _has_tools: gate for symmetry with the rest of the block. (Already flagged by Graa.)
2. string_tool_names is a list, not a set (src/cleveractors/agents/tool.py:256): tool_name not in string_tool_names is O(n). Negligible at typical tool counts (≤10); a set would also make the two checks symmetric in style.
3. Blank-line additions in tool.py (lines 94, 200): cosmetic only.

Summary

The synthesis-branch fix is correct: it builds _declared_names from self._lc_tools using the OpenAI/LangChain tool-schema convention (t.get("function", {}).get("name", "")), appends a ToolMessage with the same error wording the ticket suggests, and uses continue to keep the loop alive so the LLM can recover. The tool.py refactor splits the heterogeneous self.tools list into string_tool_names and dict_tool_names, making the dispatch check explicit.

The PR does not strictly satisfy ticket AC #3 ("the guard is applied consistently in all three tool-dispatch branches"), but the PR author has provided a sound technical justification in the ticket comments for limiting the new check to the synthesis branch: the existing ToolAgent._execute_tool safeguard + the surrounding except ExecutionError handler already prevent undeclared tool execution and deliver error feedback to the LLM in the other two branches. The new check is therefore a UX refinement for clearer LLM feedback, applied where it provides the most value (the synthesis round, where the model is being explicitly prompted to recover).

Recommended follow-up (not a blocker): Reconcile the ticket with the implementation. Either:

• Update ticket #63 to reflect the narrower scope (e.g., add a note that the safeguard in tool.py already prevents the bug in the other branches, and the new check is a UX refinement), or
• Open a follow-up ticket to apply the new _declared_names guard in the main loop and budget-exhausted branches for error-message consistency.

The existing review from Graa on the PR approves the fix and concurs on the nits above.

## PR Review: !64 (Ticket #63) ### Verdict: Approve The synthesis-branch guard is implemented correctly, and the `tool.py` refactor is a reasonable clarity improvement. The PR author has explained (in a comment on ticket #63) why they did not extend the guard to the other two tool-dispatch branches: the `_execute_tool` safeguard already rejects undeclared tool names, and the surrounding `except (ExecutionError, ConfigurationError)` block in `llm.py` already converts that into a `ToolMessage` for the LLM. The new `_declared_names` check is therefore a UX refinement for clearer LLM feedback, applied where it matters most. This reasoning is technically sound and should be reflected back into the ticket (see Summary). ### Critical Issues None. ### Major Issues None. ### Minor Issues 1. **PR description overstates the `tool.py` change as a bug fix.** - **File:** `src/cleveractors/agents/tool.py`, lines 256–259 - **Problem:** The PR description says: *"the containment check `tool_name not in self.tools` always failed because `self.tools` is a heterogeneous list of strings and dicts — a string tool name like `"file_read"` would never match."* This claim is not quite accurate — the original check `tool_name not in self.tools and tool_name not in dict_tool_names` did function correctly, because Python's `in` operator returns `False` for any string-vs-dict equality, causing the `dict_tool_names` fallback to handle dict-typed tools. The author's own ticket comment ("it just needs to be adjusted to...") implicitly acknowledges this — they propose simplifying to `if tool_name not in dict_tool_names:`, but the actual PR keeps both checks. Graa's review on the PR also notes the original code worked. - **Recommendation:** The refactor is fine as a clarity/safety improvement, but the commit message and PR description should describe it as a clarity refactor (or a defensive split), not a bug fix. Otherwise a future reader may misdiagnose the original code as broken. ### Nits 1. **`_declared_names` built outside its gate in `llm.py`** (`src/cleveractors/agents/llm.py:1058-1060`): the set comprehension runs even when there are no `synth_tool_calls`. Cost is trivial; could be moved inside the `if isinstance(synth_tool_calls, list) and synth_tool_calls and _has_tools:` gate for symmetry with the rest of the block. (Already flagged by Graa.) 2. **`string_tool_names` is a `list`, not a `set`** (`src/cleveractors/agents/tool.py:256`): `tool_name not in string_tool_names` is O(n). Negligible at typical tool counts (≤10); a set would also make the two checks symmetric in style. 3. **Blank-line additions in `tool.py`** (lines 94, 200): cosmetic only. ### Summary The synthesis-branch fix is correct: it builds `_declared_names` from `self._lc_tools` using the OpenAI/LangChain tool-schema convention (`t.get("function", {}).get("name", "")`), appends a `ToolMessage` with the same error wording the ticket suggests, and uses `continue` to keep the loop alive so the LLM can recover. The `tool.py` refactor splits the heterogeneous `self.tools` list into `string_tool_names` and `dict_tool_names`, making the dispatch check explicit. The PR does not strictly satisfy ticket AC #3 ("the guard is applied consistently in all three tool-dispatch branches"), but the PR author has provided a sound technical justification in the ticket comments for limiting the new check to the synthesis branch: the existing `ToolAgent._execute_tool` safeguard + the surrounding `except ExecutionError` handler already prevent undeclared tool execution and deliver error feedback to the LLM in the other two branches. The new check is therefore a UX refinement for clearer LLM feedback, applied where it provides the most value (the synthesis round, where the model is being explicitly prompted to recover). **Recommended follow-up (not a blocker):** Reconcile the ticket with the implementation. Either: - Update ticket #63 to reflect the narrower scope (e.g., add a note that the safeguard in `tool.py` already prevents the bug in the other branches, and the new check is a UX refinement), **or** - Open a follow-up ticket to apply the new `_declared_names` guard in the main loop and budget-exhausted branches for error-message consistency. The existing review from Graa on the PR approves the fix and concurs on the nits above.

CoreRasurae force-pushed bugfix/m2-llmagent-tool-dispatch-validation from bb3decbd96 to 4642d49a86 2026-06-30 09:26:23 +00:00 Compare

CoreRasurae added a new dependency 2026-06-30 09:49:31 +00:00

#63 fix(agents): validate LLMAgent tool dispatch against declared tools list

CoreRasurae force-pushed bugfix/m2-llmagent-tool-dispatch-validation from 4642d49a86 to bcce5dc42a 2026-06-30 11:06:15 +00:00 Compare

CoreRasurae merged commit bcce5dc42a into master 2026-06-30 11:28:31 +00:00

CoreRasurae deleted branch bugfix/m2-llmagent-tool-dispatch-validation 2026-06-30 11:28:31 +00:00

Sign in to join this conversation.

Reviewers

No reviewers

hurui200320

Labels

Clear labels

  

auto/blocked-by-deps

PR blocked by an open issue dependency. Operator must close the dep (or remove the dependency link) before the merge driver can act. Auto-cleared by merge_drive when no open deps remain.

  

auto/ci-timeout

Most recent merge cycle hit CI timeout. Driver excludes this PR while last merge_cycle row is < 30 min old; label persists thereafter as visible history.

  

auto/claimed-implementer

Currently being processed by an implementer worker.

  

auto/claimed-merge

Currently being processed by the merge driver.

  

auto/claimed-reviewer

Currently being processed by a reviewer worker.

  

auto/driver-down

Merge driver heartbeat stale; pipeline halted. Closed automatically on next clean tick.

  

auto/invariant-violation

Detected master commit violating the strict merge invariant. Tracked as an issue (not a PR label); kept here for label completeness.

  

auto/last-attempt-tier-0

In-cycle escalation: most recent attempt ran at the Tier 0 slot (`tier-0`). Slot's model defined in .opencode/models/tiers.yaml.

  

auto/last-attempt-tier-1

In-cycle escalation: most recent attempt ran at the Tier 1 slot (`tier-1`). Slot's model defined in .opencode/models/tiers.yaml.

  

auto/last-attempt-tier-2

In-cycle escalation: most recent attempt ran at the Tier 2 slot (`tier-2`). Slot's model defined in .opencode/models/tiers.yaml. Gated behind IMPLEMENTER_ESCALATION_TIER2_ENABLED.

  

auto/last-attempt-tier-min

In-cycle escalation: most recent attempt ran at the Tier -1 slot (`tier-min`). Slot's model defined in .opencode/models/tiers.yaml. Suffix is ``-min`` (not ``--1``) so the Forgejo UI reads naturally.

  

Automation Tracking

Tracking issues used by the AI Automation system for agents to communicate and report.

  

auto/needs-conflict-resolution

Rebase conflict needs LLM conflict-resolver.

  

auto/needs-implementer

Failing CI needs implementer attention.

  

auto/postmortem

Documenting a driver incident or rollback.

  

auto/ready-to-merge

Reviewer has APPROVED this PR and no later REQUEST_CHANGES is outstanding. The merge driver requires this label to even consider a PR for merging. Set by the reviewer worker on APPROVE; cleared on REQUEST_CHANGES.

  

auto/restart-throttled

Train repeatedly lost master-tempo races. Driver excludes via merge_cycle until cooldown elapses; label persists as visible history.

  

auto/revert

Revert PR backing out an invariant violation. Fast-tracked through the merge driver.

  

auto/sentinel

Sentinel PR duplicated from upstream into a personal fork by tools/duplicate_prs_to_fork.py for pipeline testing. Lives only in the fork; the canonical pipeline never sees it.

  

auto/stale-inactivity

No implementer activity for N days. Flagged for human review. Auto-cleared on next push to head branch.

  

auto/unstable

Repeatedly fails on current master (>= 3 ci-fail-on-rebased-sha releases in 12 h). Excluded from driver until human triage.

  

Blocked

A ticket in a blocked state and unable to complete until some other task is completed first.

  

Bounty

$100

A bounty of $100 for any open-source contributor who provides a MR that solves this issue

  

Bounty

$1000

A bounty of $1000 for any open-source contributor who provides a MR that solves this issue

  

Bounty

$10000

A bounty of $10000 for any open-source contributor who provides a MR that solves this issue

  

Bounty

$20

A bounty of $20 for any open-source contributor who provides a MR that solves this issue

  

Bounty

$2000

A bounty of $2000 for any open-source contributor who provides a MR that solves this issue

  

Bounty

$250

A bounty of $250 for any open-source contributor who provides a MR that solves this issue

  

Bounty

$50

A bounty of $50 for any open-source contributor who provides a MR that solves this issue

  

Bounty

$500

A bounty of $500 for any open-source contributor who provides a MR that solves this issue

  

Bounty

$5000

A bounty of $5000 for any open-source contributor who provides a MR that solves this issue

  

Bounty

$750

A bounty of $750 for any open-source contributor who provides a MR that solves this issue

  

MoSCoW

Could have

Could have feature in order to satisfy the epic/legendary.

  

MoSCoW

Must have

Must have feature in order to satisfy the epic/legendary.

  

MoSCoW

Should have

Should have feature in order to satisfy the epic/legendary.

  

Needs Feedback

There are questions in the ticket that can not be completed until the project owner provides clarity.

  

Points

1

1 man-hours worth of work for an expert with no learning curve.

  

Points

13

13 man-hours worth of work for an expert with no learning curve.

  

Points

2

2 man-hours worth of work for an expert with no learning curve.

  

Points

21

21 man-hours worth of work for an expert with no learning curve.

  

Points

3

3 man-hours worth of work for an expert with no learning curve.

  

Points

34

34 man-hours worth of work for an expert with no learning curve.

  

Points

5

5 man-hours worth of work for an expert with no learning curve.

  

Points

55

55 man-hours worth of work for an expert with no learning curve.

  

Points

8

8 man-hours worth of work for an expert with no learning curve.

  

Points

88

88 man-hours worth of work for an expert with no learning curve.

  

Priority

Backlog

This ticket has backlogged priority and is not to be worked on yet

  

Priority

CI Blocker

Critical priority issue that blocks CI/CD pipeline and prevents PR merges

  

Priority

Critical

The priority is critical

  

Priority

High

The priority is high

  

Priority

Low

The priority is low

  

Priority

Medium

The priority is medium

  

Signed-off: Owner

When an epic or legendary is in review it must be signed off by owner, tech lead, and scrum master before being marked as completed.

  

Signed-off: Scrum Master

When an epic or legendary is in review it must be signed off by owner, tech lead, and scrum master before being marked as completed.

  

Signed-off: Tech Lead

When an epic or legendary is in review it must be signed off by owner, tech lead, and scrum master before being marked as completed.

  

Spike

A ticket for learning a tool or technology that is needed to be able to do future planning and design.

  

State

Completed

The ticket has been fully implemented, completed, and merged with the source code. This label should only be applied once a ticket is closed.

  

State

Duplicate

A ticket that represents the same content as an existing ticket.

  

State

In Progress

A ticket that is actively being developed.

  

State

In Review

A ticket that has had some code completed to implement but is waiting to pass peer review and is not yet merged in.

  

State

Paused

This ticket's work started but wasn't finished. It's on hold (likely in a feature branch) and will be resumed later, either due to a blocker or a delay.

  

State

Unverified

All new tickets start in this state. A developer may set it to show the ticket is unverified. This means we haven't agreed to work on it. It will either move to a verified state or be closed as wontdo.

  

State

Verified

The issue has been verified by a developer as legitimate. It will be worked on and verified tickets are now considered part of the backlog.

  

State

Wont Do

This ticket has been decided it wont be done. This may mean the bug has been determined to not be real (cant verify) or the feature is one we have decided we dont want to adopt.

  

Type

Automation

Any edits or discussion about the AI automated coding system.

  

Type

Bug

Something that doesnt work as intended.

  

Type

Discussion

Anytime a ticket represents a discussion about a subject and doesnt fall into one of the other categories.

  

Type

Documentation

An error or improvement needed in the documentation.

  

Type

Epic

Any first tier epic. That is, an epic which contains only issues as children and will not have sub-epics.

  

Type

Feature

Some new functionality not present.

  

Type

Legendary

A type of Epic which will contain other Epics.

  

Type

Refactor

A code change that restructures existing code without changing its external behavior.

  

Type

Support

Someone needs help using the project.

  

Type

Task

A generic task that doesnt fit into the other type categories.

  

Type

Testing

Work exclusively focusing on fixing or expanding testing.

No labels

auto/blocked-by-deps

auto/ci-timeout

auto/claimed-implementer

auto/claimed-merge

auto/claimed-reviewer

auto/driver-down

auto/invariant-violation

auto/last-attempt-tier-0

auto/last-attempt-tier-1

auto/last-attempt-tier-2

auto/last-attempt-tier-min

Automation Tracking

auto/needs-conflict-resolution

auto/needs-implementer

auto/postmortem

auto/ready-to-merge

auto/restart-throttled

auto/revert

auto/sentinel

auto/stale-inactivity

auto/unstable

Blocked

Bounty

$100

Bounty

$1000

Bounty

$10000

Bounty

$20

Bounty

$2000

Bounty

$250

Bounty

$50

Bounty

$500

Bounty

$5000

Bounty

$750

MoSCoW

Could have

MoSCoW

Must have

MoSCoW

Should have

Needs Feedback

Points

1

Points

13

Points

2

Points

21

Points

3

Points

34

Points

5

Points

55

Points

8

Points

88

Priority

Backlog

Priority

CI Blocker

Priority

Critical

Priority

High

Priority

Low

Priority

Medium

Signed-off: Owner

Signed-off: Scrum Master

Signed-off: Tech Lead

Spike

State

Completed

State

Duplicate

State

In Progress

State

In Review

State

Paused

State

Unverified

State

Verified

State

Wont Do

Type

Automation

Type

Bug

Type

Discussion

Type

Documentation

Type

Epic

Type

Feature

Type

Legendary

Type

Refactor

Type

Support

Type

Task

Type

Testing

Milestone

Clear milestone

No items

No milestone

Projects

Clear projects

No items

No project

Assignees

Clear assignees

No assignees

3 participants

Notifications

Subscribe

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Blocks

#63 fix(agents): validate LLMAgent tool dispatch against declared tools list

cleveragents/cleveractors-core

Reference

cleveragents/cleveractors-core!64

No description provided.