explore-recipes

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill reads and processes local repository files which could contain instructions intended to influence the agent's behavior.
  • Ingestion points: The workflow in SKILL.md specifies reading skills/catalog.md and various source files located under ShipSwift/SWPackage/.
  • Boundary markers: Absent; there are no instructions to the agent to treat the content of these files as untrusted or to ignore any embedded commands.
  • Capability inventory: The skill utilizes file-reading capabilities to present summaries and code details; it does not invoke tools for network communication or file modification.
  • Sanitization: Absent; the skill does not define any filtering, escaping, or validation steps for the data retrieved from the local filesystem.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 11:49 AM
Security Audit — agent-trust-hub — explore-recipes