The Agent Skills Directory

[SAFE]: The skill instructions and reference materials promote high-security standards for configuration tasks, such as enforcing restricted command allowlists and preventing the inclusion of raw secrets in configuration files. It explicitly prohibits destructive commands like recursive deletions and unauthorized privilege escalation.
[INDIRECT_PROMPT_INJECTION]: The skill involves reading and updating workspace configuration files, creating a data ingestion surface. This risk is addressed through mandatory validation stages, including schema checks and behavioral verification of effective configuration to ensure no unintended permissions are granted.
[NO_CODE]: The skill package contains no executable scripts or binaries, relying entirely on structured instructions and templates to guide agent behavior, which significantly reduces the technical attack surface.

fable-config