The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes multiple git commands (e.g., git status, git diff, git add) and repository-specific commands such as pnpm format:docs:check to validate and commit changes.
[DYNAMIC_EXECUTION]: The agent is instructed to discover and execute validation gates from local files like Makefile, Justfile, and package.json scripts. This pattern involves running arbitrary commands defined within the codebase it is operating on.
[INDIRECT_PROMPT_INJECTION]: The skill fetches data from external issue trackers and pull request metadata to automate the generation of titles and descriptions, which represents an indirect injection surface.
Ingestion points: Data from issue trackers, pull request metadata, and repository-level documentation (docs/agents/workflow/config.md).
Boundary markers: Absent; the skill does not specify delimiters or instructions to treat ingested data as untrusted.
Capability inventory: Execution of git commands, shell command execution via discovered validation gates, and network-based PR/issue modifications.
Sanitization: Absent; there is no mention of sanitizing or escaping content retrieved from external sources before it is used in prompt construction.

ziw-pr