The Agent Skills Directory

[SAFE]: The skill provides a mechanism for developers to audit their own AI implementations. It includes a scan-secrets.mjs script that performs static analysis on local files to identify hardcoded credentials, exposed environment variables, and development artifacts in build outputs. This is a standard security utility.
[SAFE]: The references/official-sources.md file serves as a directory of trusted government and regulatory domains (e.g., europa.eu, ftc.gov, ico.org.uk). It explicitly instructs the agent to verify information from these official sources rather than relying on internal memory, which is a safety best practice for compliance-related tasks.
[SAFE]: The instructions in SKILL.md emphasize risk mitigation, such as moving provider calls to server-side routes to protect API keys and implementing rate limits to prevent cost-related abuse. It also includes explicit prohibitions against making false safety claims like 'no hallucinations' or 'fully compliant'.
[SAFE]: The secret scanner script (scripts/scan-secrets.mjs) is designed with safety in mind: it redacts discovered secrets (showing only the first 4 characters and length) to prevent accidental exfiltration within analysis reports. It uses only standard Node.js modules and has no external dependencies.

ai-features-readiness