The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted data from external sources, which may contain malicious instructions designed to influence the agent's behavior.\n
Ingestion points: Data is pulled from pull request titles, descriptions, and code diffs using gh pr view, gh pr diff, and git diff as described in SKILL.md.\n
Boundary markers: The instructions do not specify the use of protective delimiters or explicit 'ignore' instructions for the external content being processed.\n
Capability inventory: The agent is authorized to use git and gh CLI tools, which provide the means to interact with the repository and project metadata.\n
Sanitization: The skill does not describe any methods for sanitizing or filtering the content retrieved from pull requests before analysis.\n- [COMMAND_EXECUTION]: The skill utilizes a defined set of shell commands to gather context and perform code reviews.\n
The allowed commands are restricted to git and gh utilities, focused on read-only operations like diff, log, and status.\n
A 'Safety Boundary' in SKILL.md explicitly prohibits the use of pipelines, scripts, and any commands that could mutate the worktree or execute project code.

biome-code-review