Home/AI Skill Platform & Ecosystem/convex-improve-convex-plugin
C

convex-improve-convex-plugin

by @convexv
4.7(120)

This Skill sends the current coding session transcript to a review endpoint for an AI post-mortem. It returns structured findings — ambiguous instructions, agent-stuck patterns, tooling failures, and wins — targeted at the runbook, bootstrap script, skills, and components rather than end-user data. Sharing is opt-in, the helper asks once and remembers the choice, and keys or tokens are redacted before upload. Run it only when the user explicitly asks, and summarize the highest-severity findings around the system.

agent-feedbacksession-transcriptpost-mortemconsentskill-improvementGitHub
Installation
npx skills add https://github.com/convex/agent-skills --skill convex-improve-convex-plugin
compare_arrows

Before / After Comparison

1 组
Before

Developers who hit friction could only reconstruct the session from memory, recalling which instructions were vague and where the agent got stuck — tens of minutes of work that still missed key details, leaving maintainers scattered feedback that was hard to reproduce.

After

This Skill captures the current session transcript and returns structured findings filed by title, target, and suggested fix, so ambiguous instructions, stuck patterns, tooling failures, and wins are all captured and ready to act on.

SKILL.md

improve-convex-plugin

Sends the current coding session transcript to the anteater POST /review endpoint for an AI post-mortem. The review returns structured findings (ambiguous instructions, agent-stuck patterns, tooling failures, wins) targeted at the runbook, bootstrap script, skills, and components — not end-user data. Sharing is opt-in: the anteater-served helper asks once (Always / Just this once / Never) and remembers the choice.

Workflow

  1. Run the anteater-served helper: curl -fsSL "<anteater>/send-transcript" | bash -s -- --idea "<one-line app idea from this session>".
  2. If it prints CONSENT_REQUIRED (exit 4), the user has not chosen yet — ask them to share Always, Just this once, or Never, then re-run appending --consent always|once|never. Do not send until they answer.
  3. Watch for output markers: REVIEW_SOURCE (transcript found), REVIEW_SUBMITTED id=... (accepted), REVIEW_DONE status=done (findings ready).
  4. Summarize the highest-severity findings for the user: title → target → suggestedFix, then wins. Keep the summary about the system, not the user's data.

Rules

  • Never send a transcript until the user has explicitly chosen to share (the helper prints CONSENT_REQUIRED and exits until they do).
  • REVIEW_NO_TRANSCRIPT means no Claude/Codex .jsonl was found — tell the user.
  • Never paste raw secrets back — the script redacts keys/tokens before upload; keep the summary system-focused.
  • This is a system-improvement loop, not end-user feature feedback.

User Reviews (0)

Write a Review

Effect
Usability
Docs
Compatibility

No reviews yet

Statistics

Installs45.4K
Rating4.7 / 5.0
Version
Updated2026年11月21日
Comparisons1

User Rating

4.7(120)
5
37%
4
43%
3
13%
2
5%
1
2%

Rate this Skill

0.0

Compatible Platforms

🤖claude-code

Timeline

Created2026年10月17日
Last Updated2026年11月21日
🎁 Agent Knowledge Cards
Survey