The Agent Skills Directory

[SAFE]: The skill automates the creation of Architecture Decision Records (ADRs) in a local docs/adr/ directory.\n- [SAFE]: It employs a validation step to ensure documentation is only created for significant, hard-to-reverse decisions, which helps prevent storage of unnecessary or malicious data.\n- [SAFE]: The agent is instructed to display the drafted content for user approval rather than committing changes automatically, maintaining human oversight.\n- [SAFE]: No network activity, sensitive data access, or external code execution is present in the instructions.\n- [INDIRECT_PROMPT_INJECTION]: The skill creates content for future AI consumption, but the risk of injection is minimized by the constrained format (1–3 sentences) and the requirement for user review.

to-adr