The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses subprocess.run within its test suite to execute its own bundled scripts for validation and regression testing.- [DATA_EXFILTRATION]: The skill reads local files and records to compute hashes and normalizations. These operations are strictly bounded to user-provided paths and no evidence of unauthorized data exposure or exfiltration was found.- [PROMPT_INJECTION]: The skill handles untrusted data which presents a risk of indirect prompt injection. It mitigates this via a non-mutating inspection phase and a mandatory approval gate for all data modifications. Ingestion occurs through JSON requests; boundaries are establishing by core contracts; capabilities include bounded file reads; and sanitization is performed via adapter-specific normalization policies.

dedupe