verification-before-completion
Evidence requirement enforcement ensuring all claims are backed by logs, test results, or exit codes. Zero = success, non-zero = failure. No guessing allowed.
Best use case
verification-before-completion is best used when you need a repeatable AI agent workflow instead of a one-off prompt.
Evidence requirement enforcement ensuring all claims are backed by logs, test results, or exit codes. Zero = success, non-zero = failure. No guessing allowed.
Teams using verification-before-completion should expect a more consistent output, faster repeated execution, less prompt rewriting.
When to use this skill
- You want a reusable workflow that can be run more than once with consistent structure.
When not to use this skill
- You only need a quick one-off answer and do not need a reusable workflow.
- You cannot install or maintain the underlying files, dependencies, or repository context.
Installation
Claude Code / Cursor / Codex
Manual Installation
- Download SKILL.md from GitHub
- Place it in
.claude/skills/verification-before-completion/SKILL.mdinside your project - Restart your AI agent — it will auto-discover the skill
How verification-before-completion Compares
| Feature / Agent | verification-before-completion | Standard Approach |
|---|---|---|
| Platform Support | Not specified | Limited / Varies |
| Context Awareness | High | Baseline |
| Installation Complexity | Unknown | N/A |
Frequently Asked Questions
What does this skill do?
Evidence requirement enforcement ensuring all claims are backed by logs, test results, or exit codes. Zero = success, non-zero = failure. No guessing allowed.
Where can I find the source code?
You can find the source code on GitHub using the link provided at the top of the page.
SKILL.md Source
# Verification Before Completion ## Overview Enforces the evidence-first principle: every claim of success, fix, or completion must be backed by concrete evidence. Exit codes, test output, and logs are the only acceptable proof. ## Evidence Types 1. **Exit codes**: zero = success, non-zero = failure 2. **Test output**: pass/fail counts, coverage percentages 3. **Logs**: error messages, stack traces, resolution confirmation 4. **Build output**: compilation success/failure with timestamps ## Rules - Never claim "should work" without evidence - Never claim success without exit code 0 - Never claim a bug is fixed without reproduction failure - Always capture and report exit codes - Always run the full test suite, not just targeted tests - Use timeout guards (`timeout 60s`) to prevent hanging ## When to Use - Before completing any BUILD workflow step - Before claiming a DEBUG fix is verified - Before approving a REVIEW result - Before marking any task as done ## Agents Used - `integration-verifier` (primary consumer) - `component-builder` (TDD evidence) - `bug-investigator` (fix evidence)
Related Skills
formal-verification
Formal property verification and model checking skill for FPGA designs
requirements-verification
Skill for aerospace requirements verification and validation matrix management
shell-completion-generator
Generate shell completion scripts for bash, zsh, and fish from CLI command definitions. Creates intelligent completions with argument suggestions, file completions, and dynamic values.
verification
Verification-before-completion discipline ensuring all success criteria are met, tests pass, and reviews complete before declaring work done.
Completion Verification
Verify all phases are complete with weighted quality scoring before allowing session exit.
verification-suite
Plan structure validation, phase completeness checks, reference integrity verification, and artifact existence confirmation. Provides the structured verification layer ensuring GSD artifacts are well-formed and complete.
process-builder
Scaffold new babysitter process definitions following SDK patterns, proper structure, and best practices. Guides the 3-phase workflow from research to implementation.
babysitter
Orchestrate via @babysitter. Use this skill when asked to babysit a run, orchestrate a process or whenever it is called explicitly. (babysit, babysitter, orchestrate, orchestrate a run, workflow, etc.)
yolo
Run Babysitter autonomously with minimal manual interruption.
user-install
Install the user-level Babysitter Codex setup.
team-install
Install the team-pinned Babysitter Codex workspace setup.
retrospect
Summarize or retrospect on a completed Babysitter run.