How does this skill handle multi-agent delegation?

It instructs the orchestrator to verify the actual state of the code and VCS diffs rather than trusting the success report of a sub-agent.

Can I skip verification if the fix is simple?

No, the skill enforces an 'Iron Law' that prohibits any success claims without fresh verification, regardless of the complexity or confidence level.

What counts as valid verification evidence?

Evidence includes fresh test command outputs (0 failures), exit codes (exit 0), build logs, and line-by-line requirement checklists.

Why is evidence-based verification necessary for AI?

It prevents AI agents from making 'hallucinated' claims about code working when tests haven't actually been run or passed, ensuring a higher level of trust.

Verification Before Completion

Name: Verification Before Completion
Author: Ibrahim-3d

byIbrahim-3d

•

291

•

Seguridad y Pruebas

Enforces rigorous automated verification and evidence-based reporting before any task is marked as complete or successful.

This skill establishes a strict Gate Function for AI agents, preventing premature claims of success by mandating the fresh execution of verification commands. It addresses common failures like trusting outdated logs or assuming code fixes work without testing, ensuring that every assertion of completion is backed by tangible evidence such as exit codes, test outputs, or build logs. This is essential for maintaining high software quality and trust within multi-agent orchestration systems, as it prevents the shipping of undefined functions or incomplete requirements.

Características Principales

01Strict 'Red Flag' detection for vague language like 'should' or 'probably'

02Mandatory execution of fresh verification commands before claims

03291 GitHub stars

04Standardized Red-Green TDD cycle validation for bug fixes

05Independent verification of delegated agent tasks and VCS diffs

06Identification of specific commands required to prove success

Casos de Uso

01Ensuring local builds and linters pass before creating pull requests

02Auditing multi-agent workflows to confirm all requirements are met with evidence

03Validating bug fixes through full regression testing before committing code

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add ibrahim-3d/conductor-orchestrator-superpowers verification-before-completion

For use in Claude.ai and ChatGPT

Download Skill

Características Principales

01Strict 'Red Flag' detection for vague language like 'should' or 'probably'

02Mandatory execution of fresh verification commands before claims

03291 GitHub stars

04Standardized Red-Green TDD cycle validation for bug fixes

05Independent verification of delegated agent tasks and VCS diffs

06Identification of specific commands required to prove success

Casos de Uso

01Ensuring local builds and linters pass before creating pull requests

02Auditing multi-agent workflows to confirm all requirements are met with evidence

03Validating bug fixes through full regression testing before committing code

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add ibrahim-3d/conductor-orchestrator-superpowers verification-before-completion

For use in Claude.ai and ChatGPT

Download Skill