/audit - Code Audit Orchestrator

Orchestrate specialized review agents with confidence-based filtering.

Input

• Target scope: $1 (optional)
• If $1 is empty → select focus via AskUserQuestion, then review staged/modified files

Audit Focus

Question	Options
Focus	security / performance / readability / all

Execution

Step	Action
1	Run project static analysis tools (see Pre-flight below)
2	Task with subagent_type: audit-orchestrator (pass pre-flight results as context)
3	Orchestrator runs agents (see audit-orchestrator.md)
4	Integrator aggregates findings (Strong Inference: ≥3 root cause hypotheses → eliminate)
5	Save snapshot (see Snapshot Naming below)
6	Compare with previous snapshot, display delta
7	Output report using template

Pre-flight: Static Analysis

Auto-detect and run project lint/check tools before agents start.

Step 1: Detect task runner from project root

File	Runner
package.json	npm / yarn / pnpm / bun
composer.json	composer
Makefile	make
Taskfile.yml	task
Cargo.toml	cargo
pyproject.toml	poetry / uv / ruff
Gemfile	bundle exec

Step 2: Find lint/check scripts in detected runner

Common names: lint, typecheck, type-check, check, analyse, analyze, static, phpstan, clippy

Fallback (best-effort): If no runner found, check for config files (e.g. tsconfig.json → npx tsc --noEmit, ruff.toml → ruff check).

Step 3: Run discovered scripts

Rule	Behavior
No tools found	Skip pre-flight, proceed to agents
Non-zero exit	Capture output as context, do NOT block audit
Multiple scripts	Run independent scripts in parallel
Timeout	60s per script; kill and proceed on timeout

Snapshot Naming

SNAPSHOT="$HOME/.claude/workspace/history/audit-$(date -u +%Y-%m-%d-%H%M%S).yaml"

Example output: audit-2026-01-23-031812.yaml

Templates

Template	Purpose
@../templates/audit/output.md	Output format with delta
@../templates/audit/snapshot.yaml	Snapshot schema

Verification

Check	Required
Task called with subagent_type: audit-orchestrator?	Yes
Snapshot saved?	Yes
Delta comparison displayed?	Yes