B0yko

Proof Of Done — Development skill for Claude Code

Development community

Claude Code plugin that stops a coding agent from ending its turn with 'tests pass', 'build succeeds' or 'deployed' unless the transcript shows the command ran after the last edit and succeeded.

How to install Proof Of Done

This entry records only its repository, not the path inside it, so there is no exact command to give. Open B0yko/proof-of-done and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Proof Of Done does

Claude Code plugin that stops a coding agent from ending its turn with 'tests pass', 'build succeeds' or 'deployed' unless the transcript shows the command ran after the last edit and succeeded. Plus a CLI that audits agent transcripts for unsupported claims.

Alternatives in Development

  • Done Check — PUA Done Check — 用于没跑测试别说完成、已完成但没证据、done without proof、需要验收/回归/交付质量检查的场景 19.5k ★
  • HERO Anti OverDefense — HERO = Hashing · Edge cases · Rubrics · Overbuild — the four shapes coding agents over-defend in 333 ★
  • Claude Blocker — Block distracting websites unless Claude Code is actively running inference 282 ★

README

proof-of-done: a coding agent says 'All 42 tests pass.'; the Stop hook blocks because the last test run came before the last edit, and suggests 'uv run pytest -q'

CI Release PyPI Python 3.9–3.13 Claude Code plugin License: Apache-2.0

Quickstart · How it works · Results · Configuration · Docs


Coding agents end turns with *"all tests pass"*, *"the build succeeds"*, *"fixed"* or *"deployed"*, and the session often does not back it up. **proof-of-done** treats those words as claims. A Claude Code `Stop` hook checks the agent's own transcript: did the matching command run **after the last relevant edit**, and did it **succeed**? If not, the stop is blocked and the