skills.rest
sandsower
sandsowerCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

verify

Requires running verification commands and reading output before claiming work is complete.

10|Updated May 12, 2026
Search Tags:#code quality#testing#verification#evidence#ci#workflow gate#completion claims
One-click install
npx skills add https://github.com/sandsower/beislid --skill verify-sandsower
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: verify
Source: https://github.com/sandsower/beislid/tree/main/skills/verify
Command: npx skills add https://github.com/sandsower/beislid --skill verify-sandsower

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Agents often claim work is done, tests pass, or bugs are fixed without actually running verification, leading to false success reports. This Skill enforces a hard gate: no completion claims without fresh verification evidence. ## Core Features & Use Cases - Evidence-First Gate: Blocks any claim of completion until the verification command is run fresh, output is read, and results confirm the claim. - Claim-to-Evidence Mapping: Defines required evidence for common claims such as tests passing, linter clean, build success, and bug fixes (red to green). - Proof Requirements Matching: Aligns verification evidence with Work Contract proof_requirements like command_gate, review, ci_check, and screenshot_show_me. - Use Case: After a subagent reports a bug fix, the agent checks the VCS diff, reruns the failing test independently, and only then confirms the fix with pasted output. ## Quick Start Before telling me the fix is done, run the test suite, read the output, and show me the evidence that it passes.

Frequently Asked Questions about verify

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

FAQPage Schema
How do I verify code changes before claiming they work?▼

Identify the relevant verification command such as the test suite, linter, or build, run it fresh rather than relying on memory, read the actual output, and confirm it supports your claim before stating anything is done.

What evidence is required to claim tests pass or a bug is fixed?▼

Claiming tests pass requires test command output showing zero failures. Claiming a bug is fixed requires a previously failing test now passing, demonstrating a red-to-green transition with actual command output.

Can I trust a subagent's success report as verification?▼

No, agent success reports are not evidence. After a subagent completes, check the VCS diff, run verification independently, and only then confirm the work based on your own observed output.

What phrases indicate unverified completion claims?▼

Phrases like "should work", "probably works", "seems to be working", or "this should resolve" signal guessing without evidence. Replace them by running the command, pasting the output, and stating the observed fact.

When should visual proof be used instead of command output?▼

When verification is clearer as screenshots, videos, diffs, or a browsable HTML report, suggest the show-me skill. It only runs on explicit user request and is never triggered automatically.

Related Skills

Explore More Software Engineering Skills→
ZenkakuHiraganaZenkakuHiragana

implementation

Executes repository changes from a normalized task contract with validation.

Community
Advanced
vTRKAvTRKA

pre-pr-check

Verify pull request diffs against intent and produce evidence bundles.

Community
Advanced
openaiopenai
488

report-to-google-slides

Converts an HTML analytics report into an editable PPTX deck for Google Slides imports.

Official
Advanced
sebastianspickersebastianspicker

orchestration-quality-static

Enforce linting, formatting, and type-checking as a mandatory quality gate.

Community
Intermediate