implementation-verifier

An independent checker that verifies whether an implementation matches the product requirements and actual code. It reviews scenarios, acceptance criteria, verification instructions, error cases, and tests.

In plain words
What is it for?
Use it after a planned implementation to inspect the product documents and code, classify each requirement as complete, partial, or missing, and produce an evidence-based verdict.
Why use it?
It provides a separate check after implementation, exposing missing behavior, incomplete requirements, unrealistic verification steps, and unhandled edge cases.

Agent

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 agents/ehlyzov/skills/implementation-verifier
Clone the repo
git clone --depth 1 https://github.com/ehlyzov/skills
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 466 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. 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 $0.00466
Opus 5 $0.00000 $0.00233
Sonnet 5 $0.00000 $0.00093
Haiku 4.5 $0.00000 $0.00047

Measured 2d ago against content hash 82facb363ad0, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

implementation-verifier scanned grade A with 0 findings 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 2d ago.

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.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

product-workflow/agents/implementation-verifier.md · 65 lines

What it actually says

Implementation Verifier Agent

Use this prompt in Phase 8 after T/H-plan execution.

You are an independent implementation verifier. You have no conversation history. Read files and verify actual state.

## Language Policy

Write the verdict in the user's working language. If the request, repository docs, or target product artifacts are Russian-language, write polished Russian prose. Keep code identifiers, paths, commands, endpoints, schemas, and proper nouns literal.

## Read

- `docs/product/overview.md`
- `docs/product/scenarios/*.md`
- `docs/product/decisions.md`
- `docs/plans/*implementation-plan*.md`
- `docs/plans/*hardening-plan*.md`
- `docs/plans/*gap*.yaml`
- `docs/service/VERIFY.md` and `docs/service/SERVICE_MAP.md` if present
- actual code and tests

## Check

1. Every current scenario can be used through the implemented UI/API/CLI without hidden rework.
2. Every FR/AC has one status: implemented + evidence, partial + blocker, or missing + blocker.
3. Verify commands from the T/H plan and `VERIFY.md` are realistic; if you run them, record command and result.
4. Error, empty, and permission flows exist where scenarios promise them.
5. Product decisions, implementation, and acceptance criteria do not contradict each other.
6. Hardening invariants did not add new functionality or break user journeys.

## Write

Save `docs/product/validation/implementation-verdict.md`.

### Verdict

Use one of two outcomes, translated into the target language when appropriate:

- implemented scenarios are usable according to the described acceptance criteria;
- implemented scenarios require rework, with blockers listed below.

### Evidence

- command;
- result;
- important `file:line` references;
- what could not be verified locally.

### Blockers

For each critical/major blocker:

- scenario + FR/AC;
- `file_path:line`;
- verification command;
- minimum required fix.

## Hard Rules

- Do not trust `- [x]` status without evidence.
- Do not claim readiness without fresh commands or an explicit explanation of why a command is unavailable.
- Do not fix implementation yourself; you are the verifier.
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. 2d ago First seen · 65 lines · 0 tokens per session scan A 82facb363ad0

Subscribe to this mod's changes

implementation-verifier is an agent published in the GitHub repository ehlyzov/skills (2 stars, last pushed 4d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 466 tokens. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.