Install one skill. Tell Claude to scan your app. Get findings, evidence, and AI-assisted fixes — without leaving your terminal.
claude skill install github:launchguard/launchguard-skillWorks with Claude Code, Cursor, Codex CLI, and Gemini CLI
One command adds the skill to Claude Code. It reads your project files to find deployment URLs and Supabase credentials automatically.
claude skill install github:launchguard/launchguard-skillTell Claude to scan your app. It runs an external security audit from our servers — probing every surface an attacker would reach.
/launchguard scan myapp.comClaude presents findings with severity, evidence, and impact. Then it offers to fix the issues directly in your codebase — RLS policies, key rotation, bucket permissions.
Missing RLS, open read/write, row counts
Service role keys in client JavaScript
Publicly listable buckets and files
Server functions callable without auth
Open to anonymous callers
Found via PostgREST error hints
Every finding includes evidence — actual data from the probe, not guesswork. Write tests use PostgREST's tx=rollback so nothing is persisted.
Deploy-triggered monitoring
After fixing issues, Claude offers to set up Ongoing Guard. It watches your app for deploys and re-scans automatically — you only get emailed when something breaks.
The free scan shows what an anonymous attacker sees. Pro shows what an authenticated attacker can do — catching IDOR and broken access control that anonymous scanning misses.
Follows the Agent Skills open standard
Scan your app in 40 seconds
Free security scanning. No signup. No config.