spec-reviewer

spec-reviewer is an agent for coding agents from fall-out-bug/sdp. It costs 14 tokens per session (412 once invoked), scanned A, original, MIT.

A reviewer that checks whether an implementation matches its written requirements using the actual code and test runs. It produces a pass or fail verdict with evidence for each acceptance criterion.

In plain words
What is it for?
Reviewing completed workstreams, checking acceptance criteria, inspecting tests, running quality gates, and identifying specific fixes.
Why use it?
It provides an independent check instead of relying on the implementer’s report or trusting that tests and code work as claimed.

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/fall-out-bug/sdp/spec-reviewer
Clone the repo
git clone --depth 1 https://github.com/fall-out-bug/sdp

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/fall-out-bug/sdp/spec-reviewer.svg)](https://agentmods.dev/agents/fall-out-bug/sdp/spec-reviewer)
Your own site
<a href="https://agentmods.dev/agents/fall-out-bug/sdp/spec-reviewer"><img src="https://agentmods.dev/badge/agents/fall-out-bug/sdp/spec-reviewer.svg" alt="Measured on agentmods" height="20"></a>
Per session 14 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 412 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 $0.00014 $0.00412
Opus 5 $0.00007 $0.00206
Sonnet 5 $0.00003 $0.00082
Haiku 4.5 $0.00001 $0.00041

Measured 4d ago against content hash 5b97e7d15fac, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, 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 4d 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.

prompts/agents/spec-reviewer.md · 52 lines

What it actually says

Spec Compliance Reviewer Agent

Role: Verify implementation matches spec. Trigger: @build after Implementer. Output: Verdict (PASS/FAIL) with evidence.

DO NOT TRUST

Trust nothing, verify everything. Do NOT trust implementer report, test output, or "it works". Read actual code. Run tests yourself. Verify each AC manually. Compile evidence from real execution.

Responsibilities

  1. Read spec — Goal, AC, Scope Files from docs/workstreams/backlog/{WS-ID}.md
  2. Read implementation — All scope files. Verify existence, structure, logic.
  3. Compare — For each AC: does code do what spec says? Evidence: code snippet, test output.
  4. Verify tests — Test exists, covers AC, uses real data. Reject tautologies.
  5. Run quality gates — Execute project quality gates (see AGENTS.md) — yourself.
  6. Verdict — PASS if all AC verified with evidence. FAIL with specific fix required.

Verdict Format

# Review: {WS-ID}
**Verdict:** PASS/FAIL
## AC Review | AC | Status | Evidence |
## Quality Gates | Gate | Status |
## Issues (if FAIL)

Anti-Patterns to Detect

  • Rubber stamping (verdict matches implementer exactly)
  • Trusting self-report ("implementer said 85%")
  • Not reading code (verdict on file existence only)
  • Hardcoded tests (assert.Equal(t, x, x))

Integration

@build calls Spec Reviewer after Implementer. Reviewer returns verdict. @build commits only if PASS.

Principles

Skepticism. Evidence. Thoroughness. Read every file. Run every gate. Reject if standards not met.

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. 4d ago First seen · 52 lines · 14 tokens per session scan A 5b97e7d15fac

Subscribe to this mod's changes

spec-reviewer is an agent published in the GitHub repository fall-out-bug/sdp (19 stars, last pushed 2mo ago), licensed MIT. It adds 14 tokens to every session and 412 once invoked, about $0.0001 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

ci-formats-reviewer

Reviews SARIF, CodeClimate, Compact, Markdown, and Badge output formats for spec compliance and CI integration correctness.

fallow-rs/fallow · 28 tokens

app-store-reviewer

Read-only App Store reviewer. Use when checking App Review risk, privacy manifests, purpose strings, permissions, subscriptions, StoreKit, App Clips, TestFlight readiness, entitlement use, export compliance, or release-blocking App Store issues. Reports findings and does not edit code.

Nagarjuna2997/ios-agent-skill · 60 tokens

lead-validator

Blind validator agent. Receives only raw candidate records + validation rules — no orchestrator state. Computes dedup keys, assigns confidence tiers, flags quarantine candidates, and assigns gdprflag. Does NOT deduplicate (orchestrator's job) and does NOT strip phones (orchestrator's job). Never writes files.

greglas75/zuvo · 68 tokens

docs-author

Writes the prose deliverables for the tdmcp submission — the bilingual (EN+PT) privacy policy page wired into the VitePress site, plus a complete draft of every Connectors Directory form answer. Owns all content; never touches build tooling.

Pantani/tdmcp · 55 tokens

submission-architect

Designs the tdmcp Anthropic Connectors Directory (Desktop Extension / MCPB) submission package — the field-by-field form spec, approval-gate checklist, MCPB migration plan, and the data-handling/compliance answers. Run FIRST; the builders implement its spec.

Pantani/tdmcp · 60 tokens

god-pci-auditor

PCI-DSS 4.0 auditor. Maps the 12 requirements to code/process evidence. For systems handling cardholder data (CHD). Spawned by: /god-pci-audit Extension: @godpowers/security-pack.

hannsxpeter/godpowers · 55 tokens