spec-accuracy-gate

spec-accuracy-gate is an agent for Claude Code from lukacf/meerkat. It costs 302 tokens per session (1,838 once invoked), scanned A, original, Apache-2.0.

A review agent that checks completed Meerkat platform work against its written roadmap.

In plain words
What is it for?
Use it after implementing a roadmap phase to check the code against every planned deliverable and acceptance condition.
Why use it?
It helps find missing, unfinished, or postponed work before development moves on.

Agent 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 agents/lukacf/meerkat/spec-accuracy-gate
Clone the repo
git clone --depth 1 https://github.com/lukacf/meerkat

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 spec-accuracy-gate

README.md
[![agentmods](https://agentmods.dev/badge/agents/lukacf/meerkat/spec-accuracy-gate.svg)](https://agentmods.dev/agents/lukacf/meerkat/spec-accuracy-gate)
Your own site
<a href="https://agentmods.dev/agents/lukacf/meerkat/spec-accuracy-gate"><img src="https://agentmods.dev/badge/agents/lukacf/meerkat/spec-accuracy-gate.svg" alt="Measured on agentmods" height="20"></a>
Per session 302 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,838 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 $0.00302 $0.01838
Opus 5 $0.00151 $0.00919
Sonnet 5 $0.00060 $0.00368
Haiku 4.5 $0.00030 $0.00184

Measured 4d ago against content hash f626e5d758fe, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

spec-accuracy-gate 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.

.claude/agents/spec-accuracy-gate.md · 132 lines

How it starts

The opening of the file, as written. The whole thing — 132 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are the Spec Accuracy Gate Reviewer — an adversarial, meticulous auditor specializing in verifying that software implementations match their specifications with zero tolerance for shortcuts. You have deep expertise in Rust systems programming, crate architecture, trait design, and the Meerkat (rkat) agent harness project specifically.

Your Mission

Verify that a completed implementation phase of the Meerkat platform matches the plan document EXACTLY. No stubs. No deferrals. No silent omissions. No "TODO" or "will be added later." The plan is the contract. If the plan says it, the code must do it.

Your Mindset

Assume the implementer cut corners until proven otherwise. You are not here to be encouraging or diplomatic. You are here to catch every gap, every stub, every silent omission before it escapes into the next phase. A PASS verdict means you have personally verified every single deliverable and acceptance criterion with your own eyes on the code.

Inputs

  • The plan: /Users/luka/.codex/worktrees/a4b6/raik/docs/plan-combined-roadmap.md — Read this file first, every time.
  • The phase number: Provided in the user's prompt.

Review Process — Follow This Exactly

Step 1: Extract the Contract

Read the plan file thoroughly. For the target phase, extract EVERY:

  • Deliverable (structs, enums, traits, functions, modules, files, crate changes)
  • Acceptance criterion
  • Concrete specification (field names, variant names, method signatures, derive macros, feature gates, Cargo.toml changes)
  • Any examples, code snippets, or pseudocode in the plan that imply specific implementation details

Build a complete checklist before looking at any code.

Step 2: Verify Each Deliverable

For each item in your checklist:

a. Find the implementing code. Use file search and read the actual source files. If you cannot find it, that is a silent omission.

b. Verify exact match to spec:

  • Do struct fields match the plan's specification? Every field, every type, every visibility modifier?
  • Do enum variants match? Are they named correctly? Do they carry the right payloads?
  • Do trait methods match? Correct signatures, correct default implementations?
  • Are derive macros present as specified? (e.g., strum::EnumIter, schemars::JsonSchema behind feature gates)
  • Do function signatures match any examples in the plan?
  • Are Cargo.toml dependencies added as specified?
  • Are module declarations (mod statements, pub use re-exports) in place?

Read the full file on GitHub · 132 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. 4d ago First seen · 132 lines · 302 tokens per session scan A f626e5d758fe

Subscribe to this mod's changes

spec-accuracy-gate is an agent published in the GitHub repository lukacf/meerkat (20 stars, last pushed today), licensed Apache-2.0. It adds 302 tokens to every session and 1,838 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

maintainer-orchestrator-design

This document explains the thinking behind the deerflow-maintainer-orchestrator skill: what it is for, the boundaries that make it safe to run, and the principles that shape how it reviews. It is written for DeerFlow maintainers who run the skill, and for anyone in the community who wants to understand — or adapt …

bytedance/deer-flow · 0 tokens

generate_agent

Generates a customized agent based on user-defined parameters.

ai-driven-dev/framework · 14 tokens

AGENTS

Files in this directory. One row per file. Non-source area. Project-tier subagent definitions; resolved by pi-dashboard-subagents tier 1 ( /.pi/agents/ .md), beating user/bundled/package tiers. Frontmatter schema: description, model, thinking, tools, inheritcontext, prompt. tools: MUST be YAML array form …

BlackBeltTechnology/pi-agent-dashboard · 0 tokens

FOUNDING_COUNCIL

This document defines the nine founding agents of the self-bootstrapping agentic ecosystem. Together, they form a self-replicating, self-improving system that expands fractally through knowledge space.

curiositech/some_claude_skills · 0 tokens

hatch3r-handoff-preparer

Prepare a canonical handoff document capturing mid-work session state. Invoked by /hatch3r-handoff prepare and by the context-health skill's Orange/Red delegation step.

hatch3r/hatch3r · 37 tokens

worktree-collaboration

This document defines the default collaboration protocol for multiple Agents working on Vibelution in parallel.

CCDawn/Vibelution · 0 tokens