verification-before-completion

Enforces running verification commands and checking output before claiming work is complete.

Updated Jul 22, 2026
One-click install
npx skills add https://github.com/johnstegeman/pi-packages --skill verification-before-completion-johnstegeman
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: verification-before-completion
Source: https://github.com/johnstegeman/pi-packages/tree/main/packages/pi-superpowers-plus/skills/verification-before-completion
Command: npx skills add https://github.com/johnstegeman/pi-packages --skill verification-before-completion-johnstegeman

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? AI agents and developers often claim work is done, tests pass, or bugs are fixed without actually running the verification commands, leading to broken commits and false status reports. This Skill enforces an evidence-before-claims discipline so no completion statement is made without fresh command output. ## Core Features & Use Cases - Verification Gate: Requires identifying, running, and reading the full output of the proving command before any success claim, commit, push, or PR creation. - Rationalization Prevention: Lists common excuses ("should work", "linter passed", "agent said success") and the red-flag phrases that signal an unverified claim. - Workflow Integration: Coordinates with beads step tracking (claiming and closing the verify step, gating on smoke-test-approved) and hands off to code review and branch-finishing skills. - Use Case: Before running git commit after a bug fix, the agent reruns the full test suite, confirms zero failures, verifies the red-green regression cycle, and only then reports the fix as complete. ## Quick Start Ask the agent to verify the fix by running the full test suite and confirming zero failures before claiming the work is complete.

Frequently Asked Questions about verification-before-completion

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I verify work is complete before committing code?

Identify the command that proves the claim, run it fresh and in full, read the complete output including exit code and failure count, and only then state the claim with that evidence. Never commit, push, or open a PR without a passing verification run since your last source edit.

What counts as valid verification evidence for tests passing?

Valid evidence is the actual test command output showing zero failures from a fresh run in the current session. Previous runs, partial checks, linter results, or assumptions like "should pass" do not qualify as verification.

Does a passing linter mean the build succeeds?

No, a linter passing is not equivalent to a successful build. The build command itself must be executed and confirmed to exit with code 0 before claiming the build succeeds.

Can I trust an agent's success report without checking?

No, agent success reports must be independently verified by inspecting the actual diffs, command outputs, and VCS changes. Trusting a delegated agent's self-reported success without checking is listed as a common failure mode.

What happens if verification cannot complete?

If verification is blocked or stopped early, the verify step stays in progress for the next session to resume, or is marked blocked with an explanatory comment. The step is never closed without passing evidence.