The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local system commands to manage its review state and verify repository health. This includes git for branch management and diffing, grep for pattern searching, and various verification commands (e.g., npm test, typecheck, lint) specified in the repository's own profile. It also utilizes a specialized CLI tool, effect-solutions, to retrieve usage patterns for the Effect TS ecosystem.
[PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as its core function involves reading and analyzing untrusted source code and documentation from the target repository.
Ingestion points: Untrusted data enters the agent context whenever the agent reads repository files for review (referenced in SKILL.md and modes/continue.md).
Boundary markers: The skill uses a structured ledger schema (reviews/REVIEW.md) and specific pass rubrics to scope the agent's tasks, though it lacks explicit delimiters to prevent the agent from following instructions embedded in the code being reviewed.
Capability inventory: The skill can execute shell commands via repository scripts and perform file system writes to apply fixes (modes/fix.md).
Sanitization: No programmatic sanitization or filtering is applied to the content read from the repository; the agent is expected to interpret content as data rather than instructions.

review-campaign