Engine Audit — Security agent for Claude Code
The engine's periodic self-review of this project.
How to install Engine Audit
Installs to ~/.claude/agents/starshipsuperjam-engine-template-engine-audit.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/StarshipSuperjam/engine-template/HEAD/.claude/agents/engine-audit.md -o ~/.claude/agents/starshipsuperjam-engine-template-engine-audit.md Restart Claude Code, or start a new session, for it to be picked up.
What Engine Audit does
name: engine-audit description: The engine's periodic self-review of this project. It reads the project's own engine state, looks hard for anything that no longer earns its place or has drifted out of date, and reports what it finds for you to decide on — it never changes anything itself. role: audit model-tier: judgment model: opus effort: high permissions: read-only output-contract: audit-finding.v1 disallowedTools: [Edit, Write, NotebookEdit, Bash]
Mandate
You are this project's
Alternatives in Security
- Audit Deep — Deep single-subsystem audit that reasons about state over time (caches, upserts, migrations, concurrent scans) 335 ★
- Offensive Security Engineer — Red-team the verification engine 179 ★
- Project Auditor — Read-only health audit of the VibeFrame monorepo 165 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Taa l10n
Use this agent to audit i18n infrastructure (hard-coded user-facing strings), maintain the TR/EN terminology g
Cost Strategist
Use when a FinOps audit of vendor/cloud/AI spend is needed — vendor tally, waste by class, keep/kill/downgrade
Self Invariants Auditor
Read-only auditor for the Invariants lens of /self-audit — flags any place where a SKILL.md, a skills/_lib/ fi
SEO Crawl Analyst
Reads a crawl export from any tool and returns the site-level findings, ordered by how much traffic each one p
Figma Analytics
Use to audit the Figma library side of the design system: override hotspots (instances drifted from their main
Fleet Review Deep
Frontier reviewer for very hard reviews. Use only when the user names it, when the change crosses a security o
Related Skills
Reduce Comments
Cut new code comments down to what earns its place — why over what, inline over doc blocks, standard labels.
Rls Security Check
A Claude Code skill that audits Supabase and Postgres Row Level Security, then writes the exact policy to fix
AWS AI Analysis Agent
Serverless agent that reviews AWS Glue Jobs and Lambdas with Claude Haiku on Amazon Bedrock: findings with pro