CookiesForAI

verification-before-completion

Verification Before Completion

Verification Before Completion is an agent skill from Superpowers: your agent runs the command that proves a claim and reads the output before it says work is done, fixed or passing.

Agent skill by Jesse Vincent · Updated

Install Verification Before Completion

In Claude Code, install the whole Superpowers plugin from Anthropic's official marketplace

Terminal
/plugin install superpowers@claude-plugins-official

Or install only this skill with the skills CLI

Terminal
npx skills add obra/superpowers --skill verification-before-completion

Or paste this into your coding agent: Install the agent skill verification-before-completion from github.com/obra/superpowers. A skill can include scripts that run on your computer, so read its source first.

What Verification Before Completion does

Verification Before Completion is one of the skills in Superpowers, Jesse Vincent's open-source method for AI coding agents. It loads when the agent is about to say work is complete, fixed or passing, and before it commits or opens a pull request. Its rule: "No completion claims without fresh verification evidence." If the agent has not run the check in the current message, it cannot claim the check passes.

Before any claim, Verification Before Completion has the agent name the command that proves it, run it in full, read the whole output and the exit code, and only then state the result with the evidence. A table spells out what each claim needs: passing tests need test output with zero failures, a build needs exit code 0, a fixed bug needs the original symptom tested again, and a subagent's reported success needs the change confirmed in the version control diff.

The red flags in Verification Before Completion include words such as "should", "probably" and "seems to", and expressions of satisfaction like "Done" before any check has run. The rule covers paraphrases and anything that implies success, not only exact phrases. A regression test counts only after a red-green cycle: it passes, fails when the fix is reverted, and passes again when the fix is restored.

When to use Verification Before Completion

  • Your agent says the tests pass when it has not run them.
  • You delegate work to subagents and need their reports checked against the actual changes.
  • You want every commit or pull request from your agent backed by real test and build output.

Which agents Verification Before Completion works in

Jesse Vincent documents Verification Before Completion for Claude Code, Codex (app and CLI), Cursor, Gemini CLI, GitHub Copilot CLI, OpenCode, Antigravity, Devin CLI, Factory Droid, Grok Build CLI, Kimi Code, Pi, Qwen Code, Hermes Agent, Muse. The open skills CLI also installs it into 78 agents, including Claude Code, Codex, Cursor, Gemini CLI, GitHub Copilot, OpenCode (we listed it with the CLI on October 1, 2026). See where each agent looks for skills.

Superpowers is written in Claude Code's vocabulary (subagent dispatch, todos, the Skill tool) and ships tool mappings for Codex, Gemini CLI, Pi, Antigravity, Hermes Agent and Muse. Installed as the full plugin, a session-start hook loads Using Superpowers so the agent reaches for skills on its own; one skill installed with the skills CLI gets no hook and may point to sibling skills you do not have.

Verification Before Completion license

Verification Before Completion is published under MIT.

  • Mutation testing: mutation-testing is Trail of Bits' agent skill for configuring mutation testing campaigns with its mewt or muton tools, reading surviving mutants, telling equivalent mutants from real test gaps, and hunting bugs they expose. (7,328 repository stars on October 2, 2026)
  • Playwright CLI: playwright-cli is Microsoft's agent skill for driving a real browser from the terminal: open pages, click and fill by element reference, take snapshots and screenshots, mock requests and generate Playwright tests. (13,721 repository stars on October 2, 2026)
  • TDD: tdd is Matt Pocock's agent skill for test-driven development: the agent agrees which public interfaces to test with you, then works one failing test and one minimal fix at a time. (273,901 repository stars on October 2, 2026)
  • Webapp Testing: Webapp Testing is Anthropic's open-source agent skill for testing local web apps: the agent writes Python Playwright scripts that click through your app, take screenshots and read console logs. (179,326 repository stars on October 2, 2026)

More from Jesse Vincent

  • Brainstorming: Brainstorming is an agent skill from Superpowers that makes your coding agent work out what you want and agree a design with you before it writes any code. (293,976 repository stars on October 2, 2026)
  • Diagnosing Superpowers: Diagnosing Superpowers is an agent skill from Superpowers that reads a session's transcripts when work went wrong and reports what happened, with every finding tied to a file and line. (293,976 repository stars on October 2, 2026)
  • Dispatching Parallel Agents: Dispatching Parallel Agents is an agent skill from Superpowers: when several problems are independent, your agent sends a focused subagent to each at once, then checks the results together. (293,976 repository stars on October 2, 2026)
  • Executing Plans: Executing Plans is an agent skill from Superpowers in which your agent works through a written plan itself, task by task with tests, then gets one fresh review of the whole branch. (293,976 repository stars on October 2, 2026)

Lists that include verification-before-completion