Skip to content
Agentmart
Submit a tool

verify

Independently verify the full-build worktree — run the real build/test commands, check for test tampering, write build/11-verify.md, and pass or fail the cycle.

Skill by Kshintak

IndexedMIT
Write a review

Chalk Labs

Need help deploying verify?

Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.

Book a call

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/kshintak-verify

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Independently verify the full-build worktree — run the real build/test commands, check for test tampering, write build/11-verify.md, and pass or fail the cycle. Use when STATUS phase is verify or the human says verify, prove it, or is it done.

Use cases: testing, writing

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md and code. Checked Nov 1, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/verify

What's inside

Skill name
verify
Description
Independently verify the full-build worktree — run the real build/test commands, check for test tampering, write build/11-verify.md, and pass or fail the cycle. Use when STATUS phase is verify or the human says verify, prove it, or is it done.
License
MIT
Instructions
31 lines in SKILL.md · view on GitHub

Files (1)

  • SKILL.md

SKILL.md (first lines)

# verify

You are not the implementer. You do not trust the implementer's claims. You prove it or you fail it.

1. Work in the full-build worktree (card body `Workspace:` line). Note `git rev-parse HEAD`.
2. Commands, in order:
   - the `verify:` command from `build/08-sandbox-plan.md`;
   - `hermes verify --json .` (detects bootstrap → build → test → start; use `--skip-start` for libraries and CLIs);
   - anything the acceptance criteria in `build/07-best-of-best.md` / `09-who-what.md` name as "done looks like";
   - every human decision in `build/gates/*.approved` (`note:` lines) and `build/decisions.md`: show evidence it is implemented (a test, a command, or file:line). One missing = FAIL.
   Record each as `cmd → exit N (Ns)`.
3. Tamper check (reward hacking guard): `git diff <base>...HEAD --stat -- '*test*' '*spec*'` and look for deleted tests, new `skip`/`xfail`/`.only`, loosened assertions, or snapshot rewrites without a stated reason. Any of these without justification in `build/runs/full-build-c*.md` is a FAIL.
4. Write `build/11-verify.md`. **Newest cycle on top**, older cycles kept below:

   ```
   ## Cycle N — <date>
   - result: PASS | FAIL
   - commit: <sha> on <branch>
   - commands:
     - <cmd> → exit 0 (12s)
   - failures: <what broke, file:line, first error line>
   - tamper check: clean | <finding>
   - verifier: <lane> (<vendor>)  — implementer was <lane> (<vendor>)
   ```
5. Append the build-log line.
6. PASS → `kanban_complete(summary, metadata={"verification": [...], "commit": sha})`.
   FAIL → `kanban_block(reason="VERIFY FAIL: <one line>")`. Do not complete. `hbs tick` starts the next build cycle with your file as its first input, or stops at the cycle cap for the human.

In chat mode (a human running this by hand in Claude Code / Cursor / Codex — never the Hermes orchestrator), do the same and update `build/STATUS.md` yourself. The verifier must be a different vendor than whoever built it.

From Kshintak/new-build-framework, MIT.

More from Kshintak/new-build-framework

Signals

Stars
0
Forks
0
Open issues
0
Last commit (this folder)
Sep 29, 2026 (1 month ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from Kshintak

Done for you

Rather not set it up yourself?

Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.

See how we build AI agents →Read the case studies

Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.

launches shipped
50+
launches shipped
media placements
500+
media placements
founders on every account
2
founders on every account

Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block

Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.

Get verify running in your business

Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.

Goes to the Chalk Labs founders, not the newsletter. Privacy

Need help deploying verify?

CallGet help
Cite this page

verify skill. Agent Mart. https://agentmart.dev/skills/kshintak-verify (last updated 2026-10-24)