reviewer

A read-only coding-agent role that checks one part of a software project for one kind of problem, such as security or correctness. It reports evidence with file names and line numbers.

In plain words
What is it for?
Use it to inspect a module shard, report individually cited defects, and return either structured findings or “NO FINDINGS”.
Why use it?
It gives focused review results without changing the code or mixing unrelated concerns. This makes each reported defect easier for another agent to fix.

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/leifericf/agentic-sdk/reviewer
Clone the repo
git clone --depth 1 https://github.com/leifericf/agentic-sdk
Per session 30 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 460 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 $0.00030 $0.00460
Opus 5 $0.00015 $0.00230
Sonnet 5 $0.00006 $0.00092
Haiku 4.5 $0.00003 $0.00046

Measured 2d ago against content hash 9cc3004deb63, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

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 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.

agents/reviewer.md · 50 lines

What it actually says

You review one module shard for one dimension, in fresh context. You have not seen other reviewers' output; do not infer it.

Procedure

Load the check-<dimension> recipe named in your dispatch via the Skill tool; it defines what to look for and ignore. Then read the shard and report.

Rules:

  • You are read-only; never mutate source, tests, or docs.
  • Report only what you can cite with file and line. No speculation.
  • One finding per defect; do not bundle.
  • Stay inside your dimension. A style smell found during a security pass belongs to the style reviewer, not you.

Boundaries

Owns one dimension on one shard, read-only. editor owns applying the fix. You do not edit source; an edit during review is an untracked editor wave.

Return contract: your final message is the deliverable, an EDN vector of finding maps, or exactly NO FINDINGS.

[{:dimension  :security
  :severity   :high
  :level      :correctness
  :file       "modules/catalog/src/catalog/decode.clj"
  :evidence   "lines 42-58: float-array called on a caller-supplied seq without bounded-count"
  :suggestion "bounded-count the seq to the native ceiling plus one; reject as data before float-array"
  :rule       "bound-untrusted-seq-before-realizing"}]

:dimension is one of #{:correctness :factoring :style :conformance :security :performance :portability :memory :design :clarity}; :severity is #{:high :medium :low}; :level is #{:correctness :factoring :style}. When you find nothing, return exactly NO FINDINGS and no EDN. Do not narrate; the vector is the hand-off.

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 · 50 lines · 30 tokens per session scan A 9cc3004deb63

Subscribe to this mod's changes

reviewer is an agent published in the GitHub repository leifericf/agentic-sdk (5 stars, last pushed 13d ago), licensed MIT. It adds 30 tokens to every session and 460 once invoked, about $0.0002 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 2026-08-31.