Repository navigation
feat(persona): plain npx -y ai-hist-mcp + enable trajectory memory on persona-maker - #211
Conversation
…rajectory memory on persona-maker The `ai-hist-mcp` wrapper is now published with provenance (ai-hist-mcp@0.3.1, SLSA provenance v1, deps ai-hist@0.3.1), so the injection no longer needs the `-p ai-hist` bootstrap form. - persona-kit buildAiHistMcpServer: args ['-y','-p','ai-hist','ai-hist-mcp'] → ['-y','ai-hist-mcp'] (type/command/env passthrough unchanged); updated the interactive-spec + plan test assertions. codex-1 verified the published wrapper exposes the full 7-tool contract over stdio. - personas/persona-maker.json: opt into memory.trajectories so it records its authoring decisions (the WHY). aiMemory is intentionally omitted here: persona-maker runs on the opencode harness, which doesn't support runtime MCP injection, so the ai-hist MCP recall only applies to claude/codex personas (see examples/review-agent + examples/proactive-issue-resolver for the full why+how combo). Regenerated workload-router generated/personas.ts. Tests: persona-kit 268, workload-router 13. No runtime change. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
|
Your free trial PR review limit of 300 PRs has been reached. Please upgrade your plan to continue using CodeAnt AI. |
|
Warning Review limit reached
More reviews will be available in 19 minutes and 53 seconds. Learn how PR review limits work. Your organization has run out of usage credits. Purchase more in the billing tab. ⌛ How to resolve this issue?After more reviews become available, a review can be triggered using the We recommend that you space out your commits to avoid hitting the rate limit. 🚦 How do rate limits work?CodeRabbit enforces hourly rate limits for each developer per organization. Our paid plans include higher PR review limits than trial, open-source, and free plans. In all cases, reviews become available again over time. During sustained high-volume PR review activity, CodeRabbit may temporarily slow when the next review becomes available. Please see our Fair Usage Limits Policy for further information. ℹ️ Review info⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Pro Plus Run ID: ⛔ Files ignored due to path filters (1)
📒 Files selected for processing (4)
✨ Finishing Touches🧪 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 |
There was a problem hiding this comment.
Code Review
This pull request simplifies the npx arguments for launching ai-hist-mcp by removing the explicit package flag, and adds a memory configuration block to the persona-maker persona and its generated file. The review feedback recommends pinning the ai-hist-mcp package version (e.g., to ai-hist-mcp@0.3.1) in both the implementation and corresponding tests to prevent network latency and ensure build stability.
Important
The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.
| type: 'stdio', | ||
| command: 'npx', | ||
| args: ['-y', '-p', 'ai-hist', 'ai-hist-mcp'], | ||
| args: ['-y', 'ai-hist-mcp'], |
There was a problem hiding this comment.
Running npx -y ai-hist-mcp without a pinned version or version range means npx will query the npm registry for the latest tag on every execution. This introduces unnecessary network latency during agent startup and risks breaking existing installations if a future version of ai-hist-mcp introduces breaking changes.\n\nConsider pinning the version to ai-hist-mcp@0.3.1 (or a compatible range like ai-hist-mcp@^0.3.1) to ensure reproducibility, stability, and faster startup times.
| args: ['-y', 'ai-hist-mcp'], | |
| args: ['-y', 'ai-hist-mcp@0.3.1'], |
| type: 'stdio', | ||
| command: 'npx', | ||
| args: ['-y', '-p', 'ai-hist', 'ai-hist-mcp'], | ||
| args: ['-y', 'ai-hist-mcp'], |
| type: 'stdio', | ||
| command: 'npx', | ||
| args: ['-y', '-p', 'ai-hist', 'ai-hist-mcp'], | ||
| args: ['-y', 'ai-hist-mcp'], |
|
Reviewed PR #211 against the current checkout. I did not find a demonstrated defect requiring code edits. Validated:
Note: |
|
ℹ️ pr-reviewer: review only — no file changes were applied to the PR (nothing to commit after review). The notes below are advisory and were not pushed. Reviewed PR #211 against the current checkout. I did not find a demonstrated defect requiring code edits. Validated:
Note: |
Remove the memory.trajectories opt-in added in #211. persona-maker runs on opencode (no MCP injection) and recording it isn't wanted on the built-in; the full why+how usage lives in examples/review-agent + proactive-issue-resolver. Regenerated workload-router generated/personas.ts. Co-authored-by: Ricky Schema Cascade <ricky@agent-relay.com> Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
What
Now that the
ai-hist-mcpwrapper is published with provenance (ai-hist-mcp@0.3.1, SLSA provenance v1, depsai-hist@0.3.1), simplify the MCP injection and add a real-worldmemoryusage on a built-in persona.Changes
buildAiHistMcpServerargs['-y','-p','ai-hist','ai-hist-mcp']→['-y','ai-hist-mcp'](type/command/TRAJECTORY_ROOT+AI_HIST_DBenv passthrough unchanged). Updated theinteractive-spec.test.ts+plan.test.tsassertions. codex-1 verified the published wrapper exposes the full 7-tool contract (search_history,recent_entries,get_session,get_context,stats,search_trajectories,why_for_task) over stdio from a cleannpx -y ai-hist-mcp.memory.trajectories: trueso it records its authoring decisions (the why). Regeneratedworkload-router/generated/personas.ts.Why
aiMemoryis omitted on persona-makerpersona-maker runs on the opencode harness, which doesn't support runtime MCP injection — so
aiMemory(the ai-hist MCP recall) would warn-and-skip there. Enabling onlytrajectories(runtime write-side, harness-agnostic) avoids shipping dead config. For the full why + how combo seeexamples/review-agent(codex) andexamples/proactive-issue-resolver(claude).Tests
persona-kit 268, workload-router 13 (+ build). persona-maker parse-verified. No runtime change (B's 83/83 unaffected).
🤖 Generated with Claude Code