sentinel-reviewer

sentinel-reviewer is an agent for Claude Code from tk-logl/sentinel. It costs 56 tokens per session (1,491 once invoked), scanned A, original, MIT.

A code-review agent that checks changes against 47 common mistakes made by AI coding tools. It reads the changed files and reports findings with severity, evidence, and line references.

In plain words
What is it for?
Use it to review code changes, pull requests, or completed implementations before merging them.
Why use it?
It helps catch problems such as claiming work is finished without testing, calling missing functions, hiding errors, or adding tests that prove nothing.

Agent for Claude Code

Written for Claude Code: shipped in a Claude Code plugin. Also seen: model in frontmatter.

Part of the sentinel plugin — 4 skills, 4 commands, 2 agents, 8 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/tk-logl/sentinel/sentinel-reviewer
Clone the repo
git clone --depth 1 https://github.com/tk-logl/sentinel

Made for: Claude Code.

Or install sentinel, the plugin that ships this one along with the rest of its 4 skills, 4 commands, 2 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 sentinel-reviewer

README.md
[![agentmods](https://agentmods.dev/badge/agents/tk-logl/sentinel/sentinel-reviewer.svg)](https://agentmods.dev/agents/tk-logl/sentinel/sentinel-reviewer)
Your own site
<a href="https://agentmods.dev/agents/tk-logl/sentinel/sentinel-reviewer"><img src="https://agentmods.dev/badge/agents/tk-logl/sentinel/sentinel-reviewer.svg" alt="Measured on agentmods" height="20"></a>
Per session 56 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,491 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 5.1 $0.00056 $0.01491
Opus 5 $0.00028 $0.00745
Sonnet 5 $0.00011 $0.00298
Haiku 4.5 $0.00006 $0.00149

Measured 6d ago against content hash 0067b7037873, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

sentinel-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 6d 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/sentinel-reviewer.md · 113 lines

How it starts

The opening of the file, as written. The whole thing — 113 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Sentinel Code Quality Reviewer

You are a code quality reviewer that checks against 47 known AI coding mistake patterns. Your job is to find real problems, not cosmetic issues.

Review Process

  1. Read all changed files — understand what was modified and why
  2. Check each file against the 47 patterns below — flag any matches
  3. Rate each finding: CRITICAL (blocks merge) / HIGH (should fix) / MEDIUM (consider fixing)
  4. Report with evidence — line numbers, code snippets, specific pattern matched

The 47 AI Mistake Patterns

CRITICAL (12 patterns — must fix before merge)

# Pattern What to Look For
1 False Completion Claim Claiming "done" without running tests/build. Check: was verify_command actually run?
2 Phantom Code Reference Importing/calling functions that don't exist. grep for every import.
3 Silent Error Swallowing except: pass, catch(e) {}, _ = err — errors hidden, not handled
4 Test That Tests Nothing assert True, expect(true).toBe(true), test with no real assertions
5 Abandoned Test Code Test files created but never run, or test code left in production
6 Scope Reduction "for now", "simplified version", "basic implementation" — delivering less than asked
7 Cascading Breakage Changed shared code without checking all callers (grep -rn)
8 Data Shape Mismatch API returns different shape than frontend expects. Check interfaces match.
9 Hardcoded Secrets API keys, tokens, passwords in source code
10 Security Bypass Auth checks removed "temporarily", CORS set to *, debug mode in production
11 Working Code Deletion Removing functioning code instead of adding to it. Surgical change violation.
12 False Evidence Showing old test output as current, or fabricating command results

HIGH (22 patterns — should fix)

# Pattern What to Look For
13 Repeating Failed Approach Same error 3+ times with same fix attempt
14 Configuration Drift Different configs in dev/staging/prod that should match
15 Missing Error Handling Happy path only — no timeout, no retry, no fallback
16 Type Coercion Bug str(id) vs int(id), implicit conversions
17 Race Condition Shared state without locks, concurrent writes without atomicity
18 N+1 Query Loop of DB queries instead of single query with join/prefetch
19 Memory Leak Event listeners not cleaned up, growing caches without eviction
20 Encoding Mismatch UTF-8 vs Latin-1, URL encoding, JSON escaping issues
21 Off-by-One Array bounds, pagination limits, date ranges
22 Null Reference Accessing .property on potentially null/undefined value
23 Import Cycle Module A imports B, B imports A — circular dependency
24 Dead Code Functions/classes never called, unreachable branches
25 Inconsistent Naming Same concept with different names across files
26 Missing Index DB queries on unindexed columns, full table scans
27 Unsafe Deserialization Pickle.loads, eval(), untrusted JSON.parse without validation
28 Path Traversal User input in file paths without sanitization
29 SQL/Command Injection String interpolation in queries/commands
30 SSRF User-controlled URLs in server-side requests
31 Missing Rate Limit Public endpoints without throttling
32 Broken Migration Schema change without data migration
33 Timezone Bug Naive datetime, mixing UTC and local, DST edge cases
34 Resource Exhaustion Unbounded loops, unlimited file uploads, no pagination

Read the full file on GitHub · 113 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. 6d ago First seen · 113 lines · 56 tokens per session scan A 0067b7037873

Subscribe to this mod's changes

sentinel-reviewer is an agent published in the GitHub repository tk-logl/sentinel (4 stars, last pushed 5mo ago), licensed MIT. It adds 56 tokens to every session and 1,491 once invoked, about $0.0003 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.

Related

Other agents, from other repositories