Why are words like 'should' or 'probably' flagged?

These words indicate a lack of empirical evidence and a reliance on assumption rather than verification, which violates the core principle of 'evidence before assertions.'

When should this skill be applied?

It should be applied before any variation of success claims, expressions of satisfaction, committing code, creating pull requests, or moving to the next task.

How does this skill handle agent delegation?

It requires the user to independently verify changes and check VCS diffs after an agent reports success, rather than trusting the agent's internal success report.

What is the primary rule of the Verified Task Completion skill?

The core 'Iron Law' is that no completion claims can be made without fresh verification evidence. If a verification command hasn't been run in the current context, the task cannot be claimed as passed.

What constitutes valid evidence for a test pass?

Valid evidence requires the actual output of a test command showing zero failures and the specific count of passing tests, not just a previous run or a partial check.

Verified Task Completion

Name: Verified Task Completion
Author: GoCodeAlone

byGoCodeAlone

0•

セキュリティとテスト

Enforces rigorous, evidence-based verification cycles before allowing any claims of task completion or code readiness.

This skill implements a strict 'iron law' of software development by requiring Claude to execute and verify command outputs before asserting that work is finished, fixed, or passing. It eliminates guesswork and false positives by mandating a fresh verification process—including tests, builds, and linters—ensuring every claim of success is backed by empirical evidence. By preventing premature PRs and broken commits, this skill maintains high-quality standards and builds trust in agentic development workflows where 'should work' is never an acceptable substitute for 'proven to work.'

主な機能

010 GitHub stars

02Independent verification of delegated sub-agent tasks

03Mandatory command execution before any success assertions

04TDD Red-Green cycle verification for regression testing

05Strict validation of exit codes for builds and linters

06Flagging of non-committal language like 'should' or 'probably'

ユースケース

01Auditing multi-step requirement checklists to guarantee full feature compliance

02Validating bug fixes by verifying the failure of the original symptom before confirming the fix

03Ensuring zero-error linter and build status before creating pull requests

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add gocodealone/claude-superpowers verification-before-completion

For use in Claude.ai and ChatGPT

主な機能

010 GitHub stars

02Independent verification of delegated sub-agent tasks

03Mandatory command execution before any success assertions

04TDD Red-Green cycle verification for regression testing

05Strict validation of exit codes for builds and linters

06Flagging of non-committal language like 'should' or 'probably'

ユースケース

01Auditing multi-step requirement checklists to guarantee full feature compliance

02Validating bug fixes by verifying the failure of the original symptom before confirming the fix

03Ensuring zero-error linter and build status before creating pull requests

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add gocodealone/claude-superpowers verification-before-completion

For use in Claude.ai and ChatGPT