The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted external data (code from feature branches and pull requests) and instructs the agent to perform actions that could execute code contained within that data.
Ingestion points: The skill reviews code from external branches and pull requests as described in SKILL.md and references/overview.md.
Boundary markers: Absent; there are no instructions for the agent to ignore or delimit instructions found within the code being reviewed.
Capability inventory: The skill instructs the agent to execute shell commands including pnpm install and pnpm test on the external code (SKILL.md).
Sanitization: Absent; the skill does not provide methods for sanitizing or validating external code before execution.
[COMMAND_EXECUTION]: The skill instructions include a bash workflow for setting up a local environment and running tests on external code.
Evidence: SKILL.md contains a "Code Review Process" section with commands: git checkout feature-branch, pnpm install, cp .env.example .env, pnpm test, and pnpm lint.

code-review