skills.rest
Sma1lboy
Sma1lboyCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

principle-prove-it-works

Verify completed work by directly inspecting the real artifact instead of proxies or self-reports.

136|8|Updated May 9, 2026
Search Tags:#testing#quality assurance#code review#validation#verification#delegation#git diff
One-click install
npx skills add https://github.com/Sma1lboy/rove --skill principle-prove-it-works-sma1lboy
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: principle-prove-it-works
Source: https://github.com/Sma1lboy/rove/tree/main/.agents/skills/pstack/skills/principle-prove-it-works
Command: npx skills add https://github.com/Sma1lboy/rove --skill principle-prove-it-works-sma1lboy

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Work declared done without direct verification has unknown correctness. Indirect signals like file timestamps, cached output, successful compilation, or an agent's own summary can all be wrong, and acting on a wrong inference costs far more than checking the source. ## Core Features & Use Cases - Direct Verification Discipline: After any task, check the real thing: run the feature, read the actual value, inspect the git diff, and confirm data flows from input to output. - Delegation Auditing: When verifying delegated or agent-produced work, inspect the actual output artifact (diff, file contents, runtime behavior) rather than trusting the delegate's summary. - Scripted Proof: Write a deterministic script that re-runs the same comparison so a reviewer can re-run it instead of trusting a one-time eyeball, keeping the artifact visible for the human. - Use Case: After an agent claims it fixed a data pipeline bug, run the pipeline end-to-end and diff the old and new compiled output with a script rather than accepting the agent's report. ## Quick Start After finishing a task, ask how to prove it actually works, then run the feature or inspect the real artifact directly before declaring it done.

Frequently Asked Questions about principle-prove-it-works

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

FAQPage Schema
How do I verify that a coding task is actually done?▼

Verify by checking the real artifact directly: build the code, run it, exercise the actual feature path, and confirm data flows from input to output. Compilation success or a passing self-report is necessary but not sufficient proof.

How to verify work delegated to an AI agent?▼

Trust artifacts, not self-reports. Inspect the actual output such as the git diff, file contents, or runtime behavior rather than the agent's summary, since agents report what they intended, not always what happened.

Why is indirect verification of task output unreliable?▼

Indirect signals like file mtimes, output freshness, cached screenshots, or derived state can be stale or wrong while appearing correct. Acting on a wrong inference costs far more than directly observing the source.

What is the strongest way to prove a change works?▼

The strongest proof is a deterministic script that re-runs the same comparison, such as diffing old and new compiled output. Keep the script's output visible as an artifact a reviewer can re-run instead of trusting your word.

When should verification artifacts be committed to the repository?▼

Commit verification artifacts only for large or complex work where the trail must be auditable later, such as a big port or migration. Most work only needs the artifact visible for the human, not committed.

Related Skills

Explore More Software Engineering Skills→
mmdmcymmdmcy

principle-prove-it-works

Verify completed work by directly inspecting the real artifact instead of proxies or self-reports.

Community
Basic
SPURSGOSPURSGO

verification-before-completion

Execute validation commands and attach outputs as completion evidence.

Community
Basic
Buckeyes22Buckeyes22

houseclean

Audit cross-drive storage and generate a prioritized migration action plan.

Community
Advanced
pluginagentmarketplacepluginagentmarketplace
3

automated-testing

Run Python-based prompt-injection and adversarial security tests in CI/CD pipelines.

Community
Intermediate