spec-reviewer

spec-reviewer is an agent for Claude Code from povvo/claudikins-kernel. It costs 291 tokens per session (1,968 once invoked), scanned A, original, MIT.

A review agent that checks whether an implementation follows its written task and acceptance criteria.

In plain words
What is it for?
Use it after a task is implemented to compare the task description, acceptance criteria, and code changes, citing file and line evidence.
Why use it?
It helps catch missing requirements, extra unrequested changes, and incorrect output before a separate code-quality review.

Agent for Claude Code

Written for Claude Code: background in frontmatter. Also seen: model in frontmatter; names the TodoWrite tool.

Part of the claudikins-kernel plugin — 4 skills, 4 commands, 7 agents, 8 hooks shipped together

Good fit Use it after a task is implemented to compare the task description, acceptance criteria, and code changes, citing file and line evidence.

Compare 6 agents from other repositories ↓
Install with agentmods
npx agentmods add agents/povvo/claudikins-kernel/spec-reviewer
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.

Clone the repo
git clone --depth 1 https://github.com/povvo/claudikins-kernel

Made for: Claude Code.

Or install claudikins-kernel, the plugin that ships this one along with the rest of its 4 skills, 4 commands, 7 agents, 8 hooks.

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 spec-reviewer

README.md
[![agentmods](https://agentmods.dev/badge/agents/povvo/claudikins-kernel/spec-reviewer.svg)](https://agentmods.dev/agents/povvo/claudikins-kernel/spec-reviewer)
Your own site
<a href="https://agentmods.dev/agents/povvo/claudikins-kernel/spec-reviewer"><img src="https://agentmods.dev/badge/agents/povvo/claudikins-kernel/spec-reviewer.svg" alt="Measured on agentmods" height="20"></a>
Per session 291 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,968 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
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 5.1 $0.00291 $0.01968
Opus 5 $0.00146 $0.00984
Sonnet 5 $0.00058 $0.00394
Haiku 4.5 $0.00029 $0.00197

Measured 8d ago against content hash 3216bbec445b, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-08, from the pricing page.

Security

Grade A, and why

spec-reviewer 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 8d 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.

agents/spec-reviewer.md · 335 lines

How it starts

The opening of the file, as written. The whole thing — 335 lines — stays where its author put it; the contents beside it link to each section on GitHub.

spec-reviewer

You verify SPEC COMPLIANCE only. "Did it do what was asked?"

Your Job

Check requirements, not quality. Code quality is code-reviewer's job.

Input

You will receive:

  1. Task description - What was supposed to be implemented
  2. Acceptance criteria - Measurable requirements
  3. Implementation diff - What was actually changed

Core Principle

Evidence-based verification. Every criterion needs a file:line reference proving it's met.

What You Check

  • Did the implementation address ALL acceptance criteria?
  • Is there any scope creep (features not in spec)?
  • Is anything missing from the requirements?
  • Does the output format match expectations?

What You DON'T Check

  • Code quality (that's code-reviewer's job)
  • Error handling quality
  • Naming conventions
  • Performance
  • Security (unless explicitly in acceptance criteria)

Verification Process

Step 1: Parse Criteria

Extract each acceptance criterion as a discrete checkable item:

Original: "Returns 401 for invalid token and 403 for expired token"

Parsed:
- Criterion 1: Returns 401 for invalid token
- Criterion 2: Returns 403 for expired token

Step 2: Locate Evidence

For each criterion, find evidence in the code:

Criterion Evidence Verdict
Returns 401 for invalid src/auth.ts:45 - throws UnauthorizedError MET
Returns 403 for expired src/auth.ts:52 - throws ForbiddenError MET

Step 3: Detect Scope Creep

Look for additions not in the spec:

Spec: "Add auth middleware"

Found:
- Auth middleware (EXPECTED)
- Rate limiting (NOT IN SPEC - scope creep)
- Logging improvements (NOT IN SPEC - scope creep)

Minor scope creep (1-2 lines, obvious necessity): Note but don't fail. Major scope creep (new features, significant additions): FAIL with explanation.

Step 4: Check Completeness

Read the full file on GitHub · 335 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. 8d ago First seen · 335 lines · 291 tokens per session scan A 3216bbec445b

Subscribe to this mod's changes

spec-reviewer is an agent published in the GitHub repository povvo/claudikins-kernel (126 stars, last pushed 4mo ago), licensed MIT. It adds 291 tokens to every session and 1,968 once invoked, about $0.0015 per session on Opus 5. 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 2026-08-30.

Related

Other agents, from other repositories

challenger

Use when: before the lead reports a root-cause conclusion, a 'done/verified' claim, an irreversible action about to run (commit/deploy/rm/push), or a 2nd-time fix — in APEX or plain conversation; also every eLicit round and Verify gate. Do NOT use for: code correctness/lint/types/API usage (sniper's job), or as a veto…

fusengine/agents · 92 tokens

code-reviewer

Use when: reviewing PRs, analyzing code quality, or checking SOLID/OWASP/Clean Code compliance. Do NOT use for: writing or implementing code (use a domain expert), or a full security penetration test (use security-auditor).

fusengine/agents · 55 tokens

security-reviewer

USE BEFORE committing security-sensitive changes (auth, crypto, routes, templates, secrets). Audits current diff for OWASP-Top-10 patterns + deps typosquatting. Read-only. Returns Critical / Important / Nice schema with file:line. Model review — not a Semgrep/CodeQL replacement.

Filip-Podstavec/claude-leverage · 67 tokens

impact-mapper

Find all call sites of a symbol with structured output on Haiku — for 'what breaks if I change X?' questions. Returns file:line:snippet triples plus tests + external risk. Bounded to prevent Opus from reading every hit inline.

Filip-Podstavec/claude-leverage · 54 tokens

code-reviewer

Code review on Sonnet — security/correctness/maintainability findings, read-only. Use before commits or PRs.

Filip-Podstavec/claude-leverage · 30 tokens

focused-reviewer

Review pre-extracted code snippets passed in the prompt — never re-reads files, never runs git diff. Hard 500-token output cap. Test of 'pass less, constrain output' design pattern against the verbose-Sonnet failure mode of code-reviewer.

Filip-Podstavec/claude-leverage · 57 tokens