feat: production hardening — V1 stabilization, error classification, perf, tools #19
Conversation
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
- Add tool-loop-guard with fingerprint-based repeat detection to prevent infinite tool call cycles in the provider boundary - Add tool-schema-compat layer for normalizing tool definitions across SDK/MCP/CLI sources into consistent function-call format - Harden auth flow with better error messages and retry UX - Add CLI model-discovery entry points - Update runtime-interception with guard integration - Strengthen proxy tool-loop with error classification - Add comprehensive tests for all new modules
- Extend ParsedError with recoverable boolean field - Add isRecoverableError() for retry decision logic - Network errors recoverable, auth/quota/model errors fatal - Unknown errors with timeout/ETIMEDOUT classified as recoverable - Add debug-level logging for failed NDJSON parse in streaming parser - Add optional errorType to ExecutionResult for executor awareness - 25 new unit tests covering all error type classifications
- Add RequestPerf class for lightweight phase timing - Instrument Bun and Node.js streaming paths with 5 markers: request:start, spawn, first-token, tool-call, request:done - Add tool execution duration logging in ToolRouter - Timing summary logged at debug level via summarize() - 6 new unit tests for perf tracker
- Add mkdir tool with recursive parent directory creation - Add rm tool with force flag for non-empty directory safety - Add stat tool returning JSON metadata (size, type, perms, timestamps) - Update all test expectations from 7 to 10 tools - Add execution tests: mkdir nested dirs, rm file/dir, stat file/dir
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.
Summary
Combined production hardening with 4 logical commits covering PRs #19-#23:
Commit 1: V1 Stabilization (PR #19/#20)
Commit 2: Error Classification (PR #21)
isRecoverableError()for retry decision logic (network=recoverable, auth/quota/model=fatal)recoverablefield onParsedErrorinterfaceerrorTypeonExecutionResultCommit 3: Performance Instrumentation (PR #22)
RequestPerfclass with 5-phase timing: request:start, spawn, first-token, tool-call, request:doneCommit 4: Tool Expansion (PR #23)
mkdir,rm,stat(7 → 10 tools)rmhas safety: refuses non-empty directories withoutforce: truestatreturns JSON metadata (size, type, permissions, timestamps)Test plan