Add AI Agent Security Mini-Audit

[MAUROCERON]

Adds an open-source, no-login self-check and launch-readiness checklist for AI agent workflows.

Why it fits:

• directly scoped to autonomous AI agent security;
• covers prompt injection, tool permissions, secrets exposure, logging, and approval gates;
• MIT-licensed and maintained in 2026;
• lightweight resource for pre-launch guardrails/compliance review.

[secxena]

Thanks for contributing. Note this duplicates #50 (same repo) — please keep a single PR per tool. The project was created only days before submission and has no adoption yet; to keep the list high-signal we hold entries to a bar of demonstrated usage and sustained maintenance. Happy to revisit as it matures.