The Agent Skills Directory

[SAFE]: The skill facilitates the configuration of security gates in CI/CD pipelines using official platform tools (harness_get, harness_update, harness_list). It appropriately references platform-managed secrets for Cosign verification and provides documentation links to official Harness resources.
[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it processes external pipeline YAML data.
Ingestion points: External pipeline configurations are ingested via harness_get as described in SKILL.md and the interactive wizard flow.
Boundary markers: Instructions do not specify the use of delimiters or 'ignore' instructions when handling the retrieved YAML context.
Capability inventory: The agent is authorized to use harness_update (to modify pipelines) and harness_list (to retrieve policy information).
Sanitization: Input data from the pipeline is mapped to new structures (e.g., repo to image_path) without explicit security validation or sanitization steps.

enforce-slsa