verification-before-completion
npx machina-cli add skill a5c-ai/babysitter/verification-before-completion --openclawVerification Before Completion
Overview
Claiming work is complete without verification is dishonesty, not efficiency.
Core principle: Evidence before claims, always.
The Iron Law
NO COMPLETION CLAIMS WITHOUT FRESH VERIFICATION EVIDENCE
Gate Function
- IDENTIFY - What command proves this claim?
- RUN - Execute the FULL command (fresh, complete)
- READ - Full output, check exit code
- VERIFY - Does output confirm the claim?
- ONLY THEN - Make the claim
Red Flags
- Using "should", "probably", "seems to"
- Expressing satisfaction before verification
- Trusting agent success reports without independent verification
Agents Used
- Process agents defined in
verification-before-completion.js
Tool Use
Invoke via babysitter process: methodologies/superpowers/verification-before-completion
Source
git clone https://github.com/a5c-ai/babysitter/blob/main/plugins/babysitter/skills/babysit/process/methodologies/superpowers/skills/verification-before-completion/SKILL.mdView on GitHub Overview
Verification-before-completion enforces evidence-first claims. It prevents declaring work done without fresh verification, promoting honesty and reliability. The core principle is evidence before claims, always, encapsulated by an iron rule: no completion claims without fresh verification evidence.
How This Skill Works
Follow the Gate Function: IDENTIFY the proving command, RUN the full, fresh command, READ the full output and exit code, VERIFY that the output confirms the claim, and ONLY THEN make the completion claim.
When to Use It
- Before committing changes to a branch.
- Before creating a pull request asserting a fix or feature is complete.
- When declaring tests pass or a bug is resolved.
- When handing work off to QA or reviewers.
- When documenting delivery to stakeholders to avoid premature claims.
Quick Start
- Step 1: IDENTIFY the proving command that would demonstrate completion.
- Step 2: RUN the full command to completion and READ its output and exit code.
- Step 3: VERIFY the results; ONLY THEN claim completion.
Best Practices
- Always use fresh verification evidence rather than relying on past runs or summaries.
- Identify the proving command before running to avoid ambiguous tests.
- Run the full command to completion, not partial checks or cached results.
- Read the full output and check the exit code for alignment with the claim.
- Only publish a completion claim if verification confirms the claim.
Example Use Cases
- A developer finishes a bug fix, runs the full test suite from a clean state, verifies all tests pass and logs are clean, then opens the PR only after verification.
- Before stating a feature is working, the developer performs end-to-end tests and confirms there are no hidden errors in the run output.
- When updating a dependency, the team re-runs the build and tests to ensure the change is stable before communicating completion.
- Prior to marking a ticket as done, the engineer ensures fresh verification evidence exists (current test results, exit codes, and logs).
- During code review, the reviewer applies the gate steps to confirm the claim of completion before endorsing the change.
