The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection because it processes untrusted data from the project codebase.
Ingestion points: Code changes, file contents, and commit history retrieved via git diff, git log, and file path analysis in SKILL.md.
Boundary markers: Absent; there are no instructions provided to the agent to treat reviewed code as data rather than instructions or to use specific delimiters.
Capability inventory: The skill utilizes shell commands including git and bun run check which could be target for exploitation if the agent obeys instructions inside the code.
Sanitization: Absent; the content is analyzed without escaping or filtering mechanism.
[COMMAND_EXECUTION]: The skill uses local shell commands to facilitate the review process.
Evidence: Instructs the agent to execute git log, git diff, and bun run check (found in SKILL.md).

code-review