How do I install the Quickstart Improve skill?

Run npx skillmds@latest add get-convex/quickstart-improve in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Quickstart Improve safe to use?

This skill has not completed SkillMD's automated safety review yet. Capability flags: makes network calls. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Quickstart Improve?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Quickstart Improve free to use?

Yes. Installing skills from SkillMD is free. This skill is licensed under Apache-2.

Who published Quickstart Improve?

get-convex (@get-convex) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Quickstart Improve by get-convex — Agent Skill

Name: Quickstart Improve
Rating: 4.666666666666667 (3 reviews)
Author: get-convex

Quickstart Improve

Backward-compatible alias for improve-convex-plugin. Send THIS Codex session's transcript to the Convex quickstart backend for an AI post-mortem that improves the whole system. TRIGGER when the user runs $quickstart-improve (the former name of $improve-convex-plugin).

get-convex ab9ea1b 1.6 KB Updated Sep 10, 2026

File contents

Send session for review ($quickstart-improve → $improve-convex-plugin)

$quickstart-improve was renamed to $improve-convex-plugin; this alias keeps the old name working for existing users. The behavior is identical: ship the current Codex session transcript to anteater's /review endpoint for an AI post-mortem that returns concrete findings to improve the runbook / bootstrap / skills. Prefer $improve-convex-plugin going forward.

Run it (QB_HARNESS=codex tells the helper to read the Codex transcript):

curl -fsSL "https://basic-anteater-667.convex.site/send-transcript" \
  | QB_HARNESS=codex bash -s -- --idea "<the user's note, or the app idea>"

Read the output exactly as in improve-convex-plugin:

REVIEW_DONE status=done → summarize for the user: overall outcome + summary, then the top findings by severity (each: title → target → suggestedFix), then the wins. Keep it about the system, never paste back secrets (the helper already redacts).
REVIEW_PENDING → submitted; the review is still running (the printed /review/<id> can be re-checked).
REVIEW_NO_TRANSCRIPT / REVIEW_TRANSCRIPT_TOO_SMALL → no Codex transcript found; tell the user.
REVIEW_UPLOAD_FAILED → the endpoint was unreachable (network/sandbox); report it.

Frequently asked questions