/06 Agent Skills
audit-plan-tasks
Agent Skill that scaffolds AUDIT.md, PLAN.md, and TASKS.md so a plan is based on what you actually checked, not on the request.
The overview
Most AI plans start from the prompt. This skill starts from the code: read it, run the commands, hit the real endpoints, then write AUDIT.md, PLAN.md, and TASKS.md from what you found.
What I built
An Agent Skill that scaffolds AUDIT.md, PLAN.md, and TASKS.md after a real audit, not before one.
In practice
What you can do with it
- 01
Scaffold AUDIT.md, PLAN.md, and TASKS.md from a real audit
- 02
Install with npx skills add trivedi-vatsal/audit-plan-tasks
- 03
Keep plans grounded in code and commands, not wishful prompts