The Agent Skills Directory

[SAFE]: The skill facilitates the creation of Architecture Decision Records using a standardized template. It utilizes restricted bash commands limited to the git binary, which prevents broader shell access.
[PROMPT_INJECTION]: The skill has an indirect prompt injection surface. 1. Ingestion points: The agent reads existing ADR files and directory conventions (e.g., docs/adr/, NNNN-*.md) to determine local project standards. 2. Boundary markers: Absent. 3. Capability inventory: The skill utilizes Read, Glob, Grep, Write, and Bash(git *) tools to interact with the repository. 4. Sanitization: Absent. The final output depends on user approval, which serves as a manual safety check.

create-adr