The Agent Skills Directory

[Sensitive Content Prevention]: The skill includes a dedicated reference for content sensitivity (references/content-sensitivity.md) that provides explicit checklists to prevent the exposure of PII, secrets, and credentials. It mandates the use of placeholders for examples, which is a standard security best practice.
[Responsible AI Guardrails]: In references/rai-guardrails.md, the skill establishes instructions for the agent to treat external content as data rather than executable instructions. This design pattern helps mitigate risks associated with processing untrusted content during documentation tasks.
[Standard Validation Tooling]: The references/validation-toolchain.md utilizes common development commands (e.g., npm run lint, npm run docs:build) to perform documentation health checks. These operations are consistent with expected development workflows and do not involve unconventional or high-risk execution patterns.
[Session and State Management]: The skill maintains session state within the local repository structure (.copilot-tracking/documentation/), following standard conventions for session persistence without utilizing external network storage for tracking.

documentation