Skip to content
Agentmart
Submit a tool

steroid-verify

The verification skill for Steroid-Workflow.

Skill by nzkbuild

IndexedMIT★ 3
Write a review

Chalk Labs

Need help deploying steroid-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/nzkbuild-steroid-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

The verification skill for Steroid-Workflow. This skill runs after the engine completes all tasks, proving the code works before allowing archival. It performs core verification by default and can run optional deep scans. Produces review.md/review.json and verify.md/verify.json as durable evidence.

Use cases: productivity

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. Checked Oct 25, 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/steroid-verify

What's inside

Skill name
steroid-verify
Description
The verification skill for Steroid-Workflow. This skill runs after the engine completes all tasks, proving the code works before allowing archival. It performs core verification by default and can run optional deep scans. Produces review.md/review.json and verify.md/verify.json as durable evidence.
License
MIT
Instructions
470 lines in SKILL.md · view on GitHub

Files (1)

  • SKILL.md

SKILL.md (first lines)

# Steroid Verify (Proof of Work)

This skill runs **after** the engine completes all tasks in `plan.md`. It proves the AI's code actually works before the feature can be archived. Without verification, the pipeline is just a promise — this skill makes it proof.

The governed live outputs for this skill are:

- `.memory/changes/<feature>/review.md`
- `.memory/changes/<feature>/review.json`
- `.memory/changes/<feature>/verify.md`
- `.memory/changes/<feature>/verify.json`
- `.memory/changes/<feature>/completion.json`

The live governed authority chain for this phase is:

- `Steroid runtime contract`
- `Steroid runtime contract`
- `Steroid runtime contract`
- `Steroid runtime contract`

Adapted from:

- **Steroid Verifier** (see `Steroid internal reference`) — Goal-backward verification, 3-level artifact checks, anti-pattern scanning
- **Steroid Spec Compliance Review** (see `Steroid internal reference`) — Independent requirement verification
- **Steroid Code Quality Review** (see `Steroid internal reference`) — Structural quality checks
- **Steroid Verification Before Completion** (see `Steroid internal reference`) — Evidence-before-claims gate

## The Iron Law

```
NO COMPLETION CLAIMS WITHOUT FRESH VERIFICATION EVIDENCE
```

Source: `Steroid internal reference` — "Claiming work is complete without verification is dishonesty, not efficiency."

If you haven't run the verification command and seen its output, you cannot claim the feature passes.

## The Circuit Breaker Mandate

All terminal commands executed by this skill MUST be wrapped in the physical Node.js circuit breaker:

```
node steroid-run.cjs '<command>'
```

For file inspection and code searches, prefer shell-free helpers like `node steroid-run.cjs fs-cat ...`, `node steroid-run.cjs fs-find ...`, and `node steroid-run.cjs fs-grep ...`. Avoid `cat`, `type`, `find`, `grep`, pipes, and redirection for routine inspection because they are brittle across hardened and Windows environments.

## Phase Gate (Physical Enforcement)

Before doing anything, run the gate check:

```
node steroid-run.cjs gate verify <feature>
```

If this command fails, STOP. The engine phase is not complete.

Then run the feature verification pre-check:

```
node steroid-run.cjs verify-feature <feature>
```

This is the core verification gate. It confirms review status, task completion, runtime checks, runs the Steroid-owned AccessLint audit when local HTML targets exist, and refreshes `ui-review.md` / `ui-review.json` whenever the latest verification evidence changes the frontend review picture. The receipt records freshness metadata so later handoff surfaces can show who refreshed the frontend verdict and which evidence triggered it.

If you want optional deep scans for runtime UI behavior, code smells, and license auditing, run:

```bash
node steroid-run.cjs verify-feature <feature> --deep
```

If you already have a deployed preview, pass it explicitly:

```bash
node steroid-run.cjs verify-feature <feature> --deep --url https://preview.example.com
```

## Two-Stage Review Gate (v5.0)

Before running the verification process, ensure the two-stage review is complete:

```bash
node steroid-run.cjs review status <feature>
```

If frontend evidence changed and you want to refresh the consolidated UI verdict without rerunning the full verification command, use:

```bash
node steroid-run.cjs review ui <feature>
```

`archive <feature>` and `report generate <feature>` will also auto-refresh stale UI review receipts before they archive or hand off a UI-intensive feature.
If the refreshed `ui-review.json` is `CONDITIONAL`, Steroid now distinguishes cautionary frontend issues from blocking frontend issues. Blocking `CONDITIONAL` cases require an explicit `node steroid-run.cjs archive <feature> --force-ui` override.

If both stages show PASS, proceed to verification. If not:

1. Run `node steroid-run.cjs review spec <feature>` — then complete Stage 1
2. Run `node steroid-run.cjs review quality <feature>` — then complete Stage 2

**The review MUST pass both stages before core verification can pass and before `verify.json` can record a passing result.**

The two-stage review produces the governed review artifacts:

- `.memory/changes/<feature>/review.md` — human-readable review report
- `.memory/changes/<feature>/review.json` — machine-readable review receipt

Source: `Steroid internal reference` — "Spec compliance first, then code quality. Never skip reviews."

## The Verification Process

### Step 1: Load Context

Read these files to understand what was built:

1. `.memory/changes/<feature>/spec.md` — The acceptance criteria (what SHOULD be true)
2. `.memory/changes/<feature>/plan.md` — The task list (what was DONE)
3. `.memory/changes/<feature>/tasks.md` — The durable live task artifact from execution
4. `.memory/changes/<feature>/execution.json` — The durable execution receipt from engine
5. `.memory/changes/<feature>/context.md` — The project context (tech stack, test infra)
6. `.memory/changes/<feature>/prompt.json` — The normalized user intent, if present

From nzkbuild/steroid-workflow, MIT.

More from nzkbuild/steroid-workflow

Signals

Stars
3
Forks
0
Open issues
0
Last commit (this folder)
Mar 26, 2026 (7 months ago)
Last release
Mar 16, 2026 (7 months ago)

From GitHub (fetched Oct 24, 2026).

Details

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

More from nzkbuild

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 steroid-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 steroid-verify?

CallGet help
Cite this page

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