agent-input-sanitizer

Runs a broad "find as many issues as you can" audit by fanning out read-only subagents over disjoint surfaces, VERIFIES every finding against the real code before trusting it, then turns the survivors into a set of small, file-disjoint parallel PRs. Activate when the user asks to "audit this", "find security/robustness/UX issues", "find dozens of problems", "do a deep review and fix everything", or any broad hunt-and-fix request that should fan out across the codebase and land as multiple PRs. Distinct from peer-review (which reviews a pending diff): this audits an entire codebase from scratch and plans the fixes.

Runs a broad, multi-dimension codebase audit by fanning out parallel read-only subagents (one per dimension x file-area), independently confirms their findings, then groups the survivors into a parallelizable multi-PR fix plan and self-critiques it before delivery. Activate when the user asks to "find dozens of issues", "audit the codebase", "find security/ robustness/testing/UX deficiencies", "use subagents to audit and confirm", "find bugs and missed opportunities", or "make a plan to fix all of these in parallel across multiple PRs". Distinct from peer-review (which reviews a pending diff) and security-review (single-pass, current branch): this audits the WHOLE codebase along several axes at once and ends in a vetted fix plan, not edits.