self-validate

self-validate is a command for Claude Code from samibs/skillfoundry. It costs 0 tokens per session (2,014 once invoked), scanned A, original, MIT.

A verification command that checks whether implemented work produces the output described by its acceptance criteria.

In plain words
What is it for?
Use it to validate recent, selected, all completed, or user-interface stories by running checks and comparing actual results with expected results.
Why use it?
It catches gaps between tests or written expectations and what the running system actually does.

Command for Claude Code

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add commands/samibs/skillfoundry/self-validate
Clone the repo
git clone --depth 1 https://github.com/samibs/skillfoundry

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for self-validate

README.md
[![agentmods](https://agentmods.dev/badge/commands/samibs/skillfoundry/self-validate.svg)](https://agentmods.dev/commands/samibs/skillfoundry/self-validate)
Your own site
<a href="https://agentmods.dev/commands/samibs/skillfoundry/self-validate"><img src="https://agentmods.dev/badge/commands/samibs/skillfoundry/self-validate.svg" alt="Measured on agentmods" height="20"></a>
Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 2,014 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 1 finding. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00000 $0.02014
Opus 5 $0.00000 $0.01007
Sonnet 5 $0.00000 $0.00403
Haiku 4.5 $0.00000 $0.00201

Measured yesterday against content hash d755b059856f, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

self-validate scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

| **API endpoint** | "When POST /auth/login, Then 200 with JWT" | `curl` against running server. Check status code + response body. |
.claude/commands/self-validate.md · 200 lines

How it starts

The opening of the file, as written. The whole thing — 200 lines — stays where its author put it; the contents beside it link to each section on GitHub.

/self-validate — Output Verification Loop

Verify implemented work actually produces its expected output. Not "did tests pass?" — "does the running code match the acceptance criteria?"


Usage

/self-validate                        Validate the most recently implemented story
/self-validate [story-file]           Validate a specific story file
/self-validate --all                  Validate all DONE stories in docs/stories/
/self-validate --ui                   Validate only frontend/UI stories (browser screenshot if MCP available)

Instructions

You are the Self-Validator. Your job is not to run tests — it is to observe whether the running system produces the exact output the acceptance criteria describe. Tests are assertions on paper. This is the real check.

The Core Loop

For each story to validate:

1. LOAD        → Read the story file, extract all ACs (Gherkin: Given/When/Then)
2. CLASSIFY    → Determine verification method per AC (see table below)
3. GENERATE    → Write the exact verification command for each AC
4. EXECUTE     → Run the command, capture stdout/stderr and exit code
5. COMPARE     → Diff actual output against expected output from AC
6. REPORT      → Pass ✓ / Fail ✗ with actual-vs-expected diff
7. FIX & RETRY → If any AC fails: route to fixer with delta, re-verify (max 3 iterations)

AC Classification and Verification Method

Story Type AC Pattern Verification Method
API endpoint "When POST /auth/login, Then 200 with JWT" curl against running server. Check status code + response body.
Business logic "When calculateTax(100), Then returns 12.5" Run the specific unit test for that function. Check output.
Database "When user record created, Then email is unique-indexed" Query the DB directly. Check schema + constraints.
UI / Frontend "When dashboard loads, Then shows 5 recent jobs" If Puppeteer/Playwright MCP available: screenshot + inspect DOM. Else: curl the page, grep for expected elements.
CLI command "When sf forge --dry-run, Then outputs IGNITE phase" Run the command, grep for expected output.
File output "When forge completes, Then audit.jsonl entry written" Check file exists, cat last line, parse JSON.
Integration "When A calls B, Then C is updated" Run end-to-end sequence, check final state.

Read the full file on GitHub · 200 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. yesterday First seen · 200 lines · 0 tokens per session scan A d755b059856f

Subscribe to this mod's changes

self-validate is a command published in the GitHub repository samibs/skillfoundry (12 stars, last pushed today), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 2,014 tokens. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.