code-recon

Builds architectural context for security audits through phased codebase reconnaissance.

6|20|Updated Mar 14, 2026
One-click install
npx skills add https://github.com/andresdefi/cryptoskills --skill code-recon-andresdefi
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: code-recon
Source: https://github.com/andresdefi/cryptoskills/tree/main/skills/code-recon
Command: npx skills add https://github.com/andresdefi/cryptoskills --skill code-recon-andresdefi

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Security auditors and code reviewers waste hours orienting themselves in unfamiliar codebases before they can find vulnerabilities. This Skill provides a structured, five-phase methodology for mapping architecture, trust boundaries, entry points, and security-critical functions before deep analysis begins. ## Core Features & Use Cases - Phased Recon Methodology: Progresses from overview and technology stack detection through architecture mapping, module deep dives, function-level analysis, and detail reconnaissance. - Trust Boundary & Entry Point Mapping: Produces concrete artifacts like endpoint inventories, data flow diagrams, and call graphs for security-critical functions. - Ready-Made Resources: Includes a recon checklist, a security question bank, a context document template, and a complete FastAPI web application walkthrough example. - Use Case: Before auditing a FastAPI e-commerce API, run the recon phases to enumerate all endpoints, identify JWT authentication weaknesses, and flag a race condition in the order creation flow as a high-risk area. ## Quick Start Ask the agent to perform a security reconnaissance of your codebase and produce a context document mapping its architecture, trust boundaries, and highest-risk areas.

Frequently Asked Questions about code-recon

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I prepare a codebase for a security audit?

Start with phased reconnaissance: identify the technology stack and dependencies, map the architecture and trust boundaries, enumerate all entry points, then analyze security-critical functions. Document findings in a structured context document before hunting for vulnerabilities.

How to map trust boundaries in a web application?

Identify where trust levels change: internet to API gateway, gateway to authenticated services, services to database, and services to third-party APIs. For each boundary, record the controls in place such as TLS, JWT validation, or query parameterization, and the associated risks.

What tools help with codebase reconnaissance for security reviews?

Use grep patterns to enumerate routes and security-sensitive functions, tree-sitter for AST analysis, pyan3 or madge for call graphs, and CodeQL for taint tracking. Semgrep, bandit, trufflehog, and gitleaks cover automated SAST and secret scanning.

Does this recon methodology work for smart contracts?

Yes, it includes smart contract specific techniques such as bytecode decompilation with panoramix, storage slot analysis with cast, and event log decoding. The question bank also covers contract-specific concerns like reentrancy, oracle manipulation, and access control.

When should I use dynamic analysis instead of static code review?

Use dynamic analysis when static review is insufficient, such as discovering hidden API endpoints via mitmproxy traffic capture, observing actual SQL queries in debug mode, or extracting database schemas directly. Combine both approaches for complete coverage.