
This is a structured methodology for building deep context before security audits or code reviews, based on Trail of Bits practices. It walks you through five phases: quick overview recon (tech stack, dependencies), architecture mapping with trust boundaries, module analysis, entry point cataloging, and data flow tracing. The approach is shell-first with lots of grep and find commands to systematically catalog everything from HTTP routes to background jobs. It's overkill for familiar codebases but genuinely useful when you're dropped into an unfamiliar system and need to understand the attack surface before looking for vulnerabilities. The documentation is thorough with actual command examples rather than hand-waving.
npx -y skills add sendaifun/skills --skill zz-code-recon --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.