The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute shell scripts located in the scripts/ directory, specifically audit-typescript.sh, audit-lint.sh, and audit-similarity.sh. These are used to generate code quality metrics for the target repository.- [DATA_EXPOSURE]: The agent is directed to read and write files within the .frontend-review directory of the client repository. This behavior is consistent with the skill's stated purpose of tracking KPIs and generating hygiene reports.- [INDIRECT_PROMPT_INJECTION]: The skill processes data from raw/*.json files (outputs of the audit scripts). While this creates a theoretical surface for indirect prompt injection if the audit tools capture malicious content from the source code, the skill's scope is restricted to reporting on metrics (e.g., 'any' counts, lint errors), which limits the impact.

frontend-review-hygiene