fix: [AI-678] add stub tool definitions for historical tool_use blocks#703
Merged
anandgupta42 merged 3 commits intomainfrom Apr 13, 2026
Merged
fix: [AI-678] add stub tool definitions for historical tool_use blocks#703anandgupta42 merged 3 commits intomainfrom
anandgupta42 merged 3 commits intomainfrom
Conversation
The Anthropic API requires every `tool_use` block in message history to have a matching tool definition. When agents switch (Plan→Builder), MCP tools disconnect, or tools are filtered by permissions, the history may reference tools absent from the current set — causing a 400 error: "Requests with 'tool_use' and 'tool_result' blocks must include tool definition." Replace the LiteLLM-only `_noop` workaround with a general fix: - Extract all tool names from `tool-call` blocks in message history - Add stub definitions for any names missing from the active tools set - Stubs return "tool no longer available" if the model attempts to call them Closes #678 Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
![claude[bot]](https://avatars.githubusercontent.com/in/1236702?s=60&v=4){kind=small}
There was a problem hiding this comment.
Claude Code Review
This repository is configured for manual code reviews. Comment @claude review to trigger a review and subscribe this PR to future pushes, or @claude review once for a one-time review.
Tip: disable this comment in your organization's Code Review settings.
📝 WalkthroughWalkthroughReplaced provider-specific noop tool injection with a generic stub-tool mechanism that extracts referenced tool names from message history and injects stub definitions for missing tools. Replaced Changes
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~20 minutes Poem
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches📝 Generate docstrings
🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
… dead code Multi-model review (GPT 5.4, Gemini 3.1 Pro) identified three issues: - `toolNamesFromMessages()` now scans both `tool-call` AND `tool-result` blocks, guarding against orphaned tool-results in compacted histories - Use `Object.hasOwn()` instead of direct property check to avoid prototype pollution edge case (`toString`, `constructor`) - Remove dead `hasToolCalls()` function and its tests — sole call site was the LiteLLM workaround deleted in the previous commit Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 2
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@packages/opencode/src/session/llm.ts`:
- Around line 287-295: The toolNamesFromMessages function currently only records
parts with type "tool-call"; update it so it also captures tool-result variants
(e.g., "tool-result" and "tool_result") by checking part.type for those values
and adding part.toolName (guarding that it exists) to the names set; modify the
loop in toolNamesFromMessages to include these additional type checks so
historical tool-result entries produce stubs just like tool-call entries.
In `@packages/opencode/test/session/llm.test.ts`:
- Around line 57-65: The test is asserting the old broken behavior by expecting
an empty set from LLM.toolNamesFromMessages for a message containing a
tool-result; update the test to expect the tool name to be returned instead
(i.e., the set should contain "bash") so LLM.toolNamesFromMessages returns new
Set(["bash"]) for the provided messages; locate the test in
packages/opencode/test/session/llm.test.ts and change the expect(...) assertion
accordingly to reflect the corrected behavior needed for the Anthropic
validation path.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Repository UI
Review profile: CHILL
Plan: Pro
Run ID: f685342b-dede-4153-8a61-b16cd17739a2
📒 Files selected for processing (2)
packages/opencode/src/session/llm.tspackages/opencode/test/session/llm.test.ts
![cubic-dev-ai[bot]](https://avatars.githubusercontent.com/in/1082092?s=60&v=4){kind=small}
There was a problem hiding this comment.
1 issue found across 2 files
Prompt for AI agents (unresolved issues)
Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.
<file name="packages/opencode/src/session/llm.ts">
<violation number="1" location="packages/opencode/src/session/llm.ts:292">
P1: `tool-result` blocks also carry `toolName` and trigger the same Anthropic validation error when no matching tool definition exists. If history trimming leaves an orphaned `tool-result` without its paired `tool-call`, this scan will miss it and the API will still return a 400. Include `part.type === "tool-result"` in the condition.</violation>
</file>
Reply with feedback, questions, or to request a fix. Tag @cubic-dev-ai to re-run a review, or fix all with cubic.
packages/opencode/src/session/llm.ts
Outdated
| for (const msg of messages) { | ||
| if (!Array.isArray(msg.content)) continue | ||
| for (const part of msg.content) { | ||
| if (part.type === "tool-call") names.add(part.toolName) |
There was a problem hiding this comment.
P1: tool-result blocks also carry toolName and trigger the same Anthropic validation error when no matching tool definition exists. If history trimming leaves an orphaned tool-result without its paired tool-call, this scan will miss it and the API will still return a 400. Include part.type === "tool-result" in the condition.
Prompt for AI agents
Check if this issue is valid — if so, understand the root cause and fix it. At packages/opencode/src/session/llm.ts, line 292:
<comment>`tool-result` blocks also carry `toolName` and trigger the same Anthropic validation error when no matching tool definition exists. If history trimming leaves an orphaned `tool-result` without its paired `tool-call`, this scan will miss it and the API will still return a 400. Include `part.type === "tool-result"` in the condition.</comment>
<file context>
@@ -276,4 +278,21 @@ export namespace LLM {
+ for (const msg of messages) {
+ if (!Array.isArray(msg.content)) continue
+ for (const part of msg.content) {
+ if (part.type === "tool-call") names.add(part.toolName)
+ }
+ }
</file context>
Suggested change
| if (part.type === "tool-call") names.add(part.toolName) | |
| if (part.type === "tool-call" || part.type === "tool-result") { | |
| names.add(part.toolName) | |
| } |
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
1 participant
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.
What does this PR do?
Fixes the Anthropic API error "Requests with 'tool_use' and 'tool_result' blocks must include tool definition" that occurs when conversation history references tools no longer in the active set (e.g., after switching agents from Plan→Builder, MCP tools disconnecting, or tools filtered by permissions).
Replaces the previous LiteLLM-only
_noopworkaround with a general fix that:tool-callblocks in message historyType of change
Issue for this PR
Closes #678
How did you verify your code works?
toolNamesFromMessages()covering: empty messages, no tool calls, extraction, dedup, tool-call vs tool-result distinctionllm.test.tspass (5 new + 10 existing)bun run ci— unit tests + marker guard)Checklist
Summary by cubic
Prevents Anthropic validation errors by adding stub tool definitions for historical
tool_useblocks when tools are no longer active. Replaces the LiteLLM-only_noopworkaround with a provider-agnostic fix for AI-678.tool-callandtool-result, then adds stubs for any missing in the active set.Object.hasOwn()for safe presence checks; removes_noopand deadhasToolCalls(); adds tests fortoolNamesFromMessages(extraction, dedup, results).Written for commit f7fdcc8. Summary will update on new commits.
Summary by CodeRabbit
Bug Fixes
Tests