What problem does it solve? Security audits and architecture reviews often fail because reviewers hunt for bugs before understanding how the system fits together, producing findings nobody can judge. This Skill builds a structured model of unfamiliar code first, recording what each function assumes, guarantees, and depends on. ## Core Features & Use Cases - Per-Function Analysis: Dispatches subagents to analyze each function, recording inputs, trust levels, preconditions, effects, and block-by-block reasoning with line citations. - Unenforced Assumption Detection: Flags assumptions marked "nothing found" where code relies on a condition nothing establishes, the most valuable input for the vulnerability-hunting phase. - Cross-Domain Support: Covers smart contracts, C/C++, decompiled firmware, and web services with domain-specific guidance on trust boundaries and black-box calls. - Use Case: Starting a smart contract audit, run the skill on the repository to produce a dossier mapping every function's invariants and couplings, then hand the unenforced assumptions and open questions to the hunting phase. ## Quick Start Run the audit-context-building skill on this codebase to map each function's assumptions, guarantees, and dependencies before I start looking for vulnerabilities.
