The Agent Skills Directory

[PROMPT_INJECTION]: The skill exposes an attack surface for indirect prompt injection through the processing of untrusted repository data and task descriptions.
Ingestion points: Processes content from ./apex/tasks/[ID].md and the output of git diff to provide context for review agents.
Boundary markers: Lacks explicit delimiters or instructions to ignore embedded commands when interpolating code diffs into sub-agent prompts.
Capability inventory: The skill can modify local files during documentation updates and execute shell commands for git operations (git add, git commit).
Sanitization: No sanitization or validation of the ingested code content is performed before it is passed to the review agents.
[COMMAND_EXECUTION]: Employs standard git CLI tools for version control management as part of its primary functionality.
Evidence: Executes git diff, git status, git add, and git commit to finalize task changes.

ship