The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted data from external security scan results and SBOM documents.
Ingestion points: Untrusted content enters the agent context via the harness_list and harness_get tools when retrieving security_issue and scs_sbom resource types (referenced in SKILL.md).
Boundary markers: The instructions lack explicit boundary markers or negative constraints to prevent the agent from obeying instructions that might be embedded within vulnerability descriptions or component metadata.
Capability inventory: The skill provides the agent with write and execute capabilities, specifically the ability to create, approve, and revoke security exemptions using harness_create and harness_execute (referenced in SKILL.md).
Sanitization: There are no specified procedures for sanitizing or validating external scanner output before it is used to inform decisions or generate reports.

security-report