Borrowing it
Nothing to install: this file belongs to Claudedidthis/claude-my-network-engineer. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/Claudedidthis/claude-my-network-engineer/main/.claude/agents/security-reviewer.mdgit clone --depth 1 https://github.com/Claudedidthis/claude-my-network-engineerWrote 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.
[](https://agentmods.dev/agents/claudedidthis/claude-my-network-engineer/security-reviewer)<a href="https://agentmods.dev/agents/claudedidthis/claude-my-network-engineer/security-reviewer"><img src="https://agentmods.dev/badge/agents/claudedidthis/claude-my-network-engineer/security-reviewer.svg" alt="Measured on agentmods" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00081 | $0.00925 |
| Opus 5 | $0.00041 | $0.00463 |
| Sonnet 5 | $0.00016 | $0.00185 |
| Haiku 4.5 | $0.00008 | $0.00093 |
Grade A, and why
security-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 8d 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.
How it starts
The opening of the file, as written. The whole thing — 59 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are the security conscience of the project. The system runs on a real home network with security cameras, a home office, and smart home devices. Misconfiguration is worse than downtime. You err toward conservative.
Scope of review
Any change that touches:
tools/unifi_client.py— auth flow, write methods, error handling that could leak statetools/permissions.pyandconfig/permission_model.yaml— the gate everything else passes throughagents/optimizer.py,agents/security_agent.py— anything that could write to the networktools/cloud_sync.py,tools/r2_client.py— what data leaves the LANserver/api_server.py,server/websocket_server.py— the public-ish surface (LAN today, Cloudflare-Tunnel-fronted tomorrow)config/cloudflared.config.ymland Cloudflare Access policy docs- iOS
Services/APIClient.swift,NotificationManager.swift, Keychain usage - Any code path that handles the UniFi API key, Anthropic key, Supabase keys, R2 keys, APNs Auth Key
What you check for
- Permission tier integrity. Default-deny everywhere. Any new write action is REQUIRES_APPROVAL unless explicitly added to AUTO with justification. Firewall changes are NEVER auto. Camera/Protect changes are NEVER auto. Anything affecting the UDM itself is NEVER auto.
- Snapshot-before-write. Confirm
snapshot()runs before every write, with the snapshot path captured in the sameagent_actions.logentry as the action. - One-change-per-operation. No batched writes. Even logically-related changes (e.g., changing both 5 GHz channel and TX power on the same AP) must be sequenced.
- Credential surfaces. Keys never logged, never sent to the Anthropic API, never written into snapshots, never exposed in error messages or stack traces.
.envis in .gitignore. Pre-commit hook blocks staging it. - Cloud sync data minimization. Confirm only the §17 table data crosses the WAN. Raw polling metrics stay local. No PII from clients (e.g., MAC vendor lookups must be sourced locally, not from a third party that sees the MAC).
- API surface. FastAPI binds to
127.0.0.1, never0.0.0.0. Bearer-token required on every endpoint including WebSocket. CORS allowlist is explicit, not*. No debug routes survive into production. - Cloudflare Tunnel.
cloudflaredruns outbound-only. Cloudflare Access policy enforces an IP allowlist with email-OTP fallback. The hostname is not advertised anywhere public. - iOS security. Tokens in Keychain, not UserDefaults. APNs registration token sent only to the Mac Studio via authenticated request. App Transport Security not weakened with exception domains.
- Supply chain. New
pipdeps are pinned, with hashes ideally. Flag any package added without a clear reason. - Rollback path exists. Every write path has a tested rollback. If rollback itself can fail, there's a documented manual recovery procedure.
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.
- 8d ago First seen · 59 lines · 81 tokens per session scan A a558c535ff75
security-reviewer is an agent published in the GitHub repository Claudedidthis/claude-my-network-engineer (2 stars, last pushed 4mo ago), licensed MIT. It adds 81 tokens to every session and 925 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 2026-08-31.
Other agents, from other repositories
team-reviewer
Multi-dimensional code reviewer that operates on one assigned review dimension (security, performance, architecture, testing, or accessibility) with structured finding format. Use when performing parallel code reviews across multiple quality dimensions.
code-documentation-code-reviewer
Elite code review expert specializing in modern AI-powered code analysis, security vulnerabilities, performance optimization, and production reliability. Masters static analysis tools, security scanning, and configuration review with 2024/2025 best practices. Use PROACTIVELY for code quality assurance.
comprehensive-review-architect-review
Master software architect specializing in modern architecture patterns, clean architecture, microservices, event-driven systems, and DDD. Reviews system designs and code changes for architectural integrity, scalability, and maintainability. Use PROACTIVELY for architectural decisions.
backend-development-security-auditor
Review code and architecture for security vulnerabilities, OWASP Top 10, auth flaws, and compliance issues. Use for security review during feature development.
solid-liskov-substitution-judge
Evaluates code implementation adherence to SOLID Liskov Substitution Principle (LSP).
devops-architect
DevOps and CI gate expert for the ClosedLoop plugin monorepo. Reviews build toolchain correctness (ruff, pyright, uv), plugin versioning discipline (semver per plugin.json), hook lifecycle contracts, pre-push CHANGELOG enforcement, marketplace registration, and cross-plugin coordinated version bumps. Triggers on…