← ClaudeAtlas

create-auditlisted

Interactive wizard for creating any type of audit — code, process, infrastructure, ecosystem, or custom. Uses extensive multi-agent discovery and deep-plan Q&A to produce audit skills, multi-AI templates, and TDMS-ready output through a structured, approval-gated workflow.
jasonmichaelbell78-creator/sonash-v0 · ★ 2 · AI & Automation · score 68
Install: claude install-skill jasonmichaelbell78-creator/sonash-v0
# Create Audit Wizard <!-- prettier-ignore-start --> **Document Version:** 3.0 **Last Updated:** 2026-03-16 **Status:** ACTIVE <!-- prettier-ignore-end --> Interactive skill that creates audit infrastructure for any domain in this repository — code quality, processes, ecosystems, infrastructure, workflows, documentation, or anything else that exists here. Produces audit skills, multi-AI templates, output directories, and TDMS-ready integration through extensive discovery and approval-gated generation. ## Critical Rules (MUST follow) 1. **Extensive discovery before writing** — MUST complete multi-agent codebase exploration (with convergence loop) and deep-plan Q&A before generating any files. No skipping ahead. 2. **Approval gate before generation** — MUST present planned structure and get explicit approval before creating files. Irreversible actions need gates. 3. **Run `/skill-audit` on every created audit** — MUST run `/skill-audit` on the generated audit as a Phase 5 validation gate, not just a post-creation suggestion. 4. **Generated audits MUST follow SKILL_STANDARDS.md** — routing table, When NOT to Use, MUST/SHOULD/MAY, compaction resilience, learning loop, invocation tracking. Reference CLAUDE.md by path, don't duplicate conventions. 5. **Generated scripts MUST follow CLAUDE.md Section 5** — error sanitization via `scripts/lib/sanitize-error.js`, path traversal checks, try/catch on file reads. Reference [CLAUDE.md](../../CLAUDE.md) Secti