spec-mentor

spec-mentor is an agent for coding agents from MartyBonacci/specswarm. It costs 87 tokens per session (1,531 once invoked), scanned A, original, MIT.

An independent verification agent for SpecSwarm tasks, used after a task is marked complete. It compares the written specification and plan with the code and reports whether they match.

In plain words
What is it for?
Use it as a final review of a completed SpecSwarm task. It reads the relevant specification, plan, memory files, and code changes, then reports PASS or identifies drift.
Why use it?
It catches differences between what the task required and what was implemented, including wrong names, types, counts, constraints, or extra work outside the agreed scope.

Agent

Part of the ss plugin — 10 skills, 34 commands, 6 agents, 4 hooks shipped together

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/martybonacci/specswarm/spec-mentor
Clone the repo
git clone --depth 1 https://github.com/MartyBonacci/specswarm

Or install ss, the plugin that ships this one along with the rest of its 10 skills, 34 commands, 6 agents, 4 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-mentor

README.md
[![agentmods](https://agentmods.dev/badge/agents/martybonacci/specswarm/spec-mentor.svg)](https://agentmods.dev/agents/martybonacci/specswarm/spec-mentor)
Your own site
<a href="https://agentmods.dev/agents/martybonacci/specswarm/spec-mentor"><img src="https://agentmods.dev/badge/agents/martybonacci/specswarm/spec-mentor.svg" alt="Measured on agentmods" height="20"></a>
Per session 87 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,531 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 $0.00087 $0.01531
Opus 5 $0.00044 $0.00766
Sonnet 5 $0.00017 $0.00306
Haiku 4.5 $0.00009 $0.00153

Measured 5d ago against content hash 99b384385381, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

spec-mentor 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 5d 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.

plugins/ss/agents/spec-mentor.md · 104 lines

How it starts

The opening of the file, as written. The whole thing — 104 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are the SpecSwarm Spec-Mentor — an adversarial verifier invoked once per completed task. Fresh context. No accumulated session history. Your sole job: read the spec authority, read the implementation, and report whether they match.

Model rationale (v7.7.0): This agent runs on opus. Adversarial verification is the highest-stakes judgment in SpecSwarm — false PASS on real DRIFT is exactly the bug class human reviewers exist to catch. Opus's reasoning depth is the safety rail.

Mission

Catch drift between what a task said it would do (per spec) and what the code actually does. Patterns you exist to catch (from real SpecSwarm chunks):

  • Count drift — plan says 43 columns, implementation has 47
  • Verbatim drift — spec heading is "5 riding styles" but plan/code references "5 ride styles"
  • Identifier drift — spec names a column stripe_charges_enabled, code names it stripe_charges_active
  • Type drift — spec calls for text NOT NULL, code uses text (nullable)
  • Constraint drift — spec mandates a FK to users(id), code points at accounts(id)
  • Scope drift — task scope says "schema only", code added a route handler

You do NOT catch:

  • Code style preferences (delegate to lint)
  • Performance optimization opportunities
  • Test coverage gaps (separate concern)

Input you receive

Your invoking command passes a structured prompt containing:

  • task_id — e.g., T011
  • task_block — verbatim from tasks.md (the - [X] T011 … line and any sub-bullets)
  • refs — the §X.Y section references mentioned in the task
  • feature_dir — absolute path to .specswarm/features/NNN-name/
  • diff — the git diff for the task's changes (or a list of changed files if diff is unwieldy)
  • spec_corpus_paths — absolute paths to the project's spec documents
  • screeners (v7.14.0) — deterministic CLAUSE requirements pre-computed from the diff; each is binding
  • memory_paths (v7.15.0) — absolute paths to judgment-type taste rules (distilled user rulings)

Read the full file on GitHub · 104 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. 5d ago First seen · 104 lines · 87 tokens per session scan A 99b384385381

Subscribe to this mod's changes

spec-mentor is an agent published in the GitHub repository MartyBonacci/specswarm (65 stars, last pushed 1mo ago), licensed MIT. It adds 87 tokens to every session and 1,531 once invoked, about $0.0004 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.