The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute git diff and file read commands to gather data for analysis. This is a standard and safe operation for development-focused skills performing code reviews.
[PROMPT_INJECTION]: The skill is exposed to indirect prompt injection through the processing of untrusted code diffs and file contents. Ingestion points: Phase 1 retrieves data via git diff and direct file system reading of modified lines. Boundary markers: The instructions do not define delimiters or specific 'ignore instructions' markers for the ingested code. Capability inventory: The skill uses git for data retrieval and performs read operations; no file-write or network-access capabilities are utilized. Sanitization: No explicit sanitization or instruction-ignoring markers are applied to the code being reviewed.

find-bugs