The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the fallow CLI for codebase analysis, including potentially destructive auto-fix operations; it specifies that the agent must use dry-runs and manual confirmation to ensure safety.
[EXTERNAL_DOWNLOADS]: The skill directs the installation of the fallow tool via npm or cargo and references external CI templates and configuration schemas from the vendor's GitHub repository.
[DATA_EXFILTRATION]: Functional features include uploading build source maps and coverage inventory to the fallow.tools cloud service for runtime analysis, which involves network communication with the vendor's infrastructure for telemetry and reporting.
[PROMPT_INJECTION]: The skill provides specific 'Agent Rules' to mitigate indirect prompt injection risks by instructing the agent to treat all project-level configuration as untrusted data and to ignore any instructions found within them.
[CREDENTIALS_UNSAFE]: The skill references using sensitive environment variables like GITLAB_TOKEN for API interactions in CI/CD pipelines, which is handled according to standard development security practices.

fallow