link-strategy
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill defines a workflow for auditing links and site structure, which necessitates ingesting data from local files or external URLs. This creates a surface for indirect prompt injection where malicious instructions embedded in the audited content could influence the agent's subsequent actions.
- Ingestion points: Files identified via the Glob tool and the content of pages linked within the site structure.
- Boundary markers: The skill lacks instructions to use delimiters or to ignore instructions found within the processed data.
- Capability inventory: The skill mentions using Glob for file discovery and testing links, implying network and file system access.
- Sanitization: There are no instructions for sanitizing or validating input data before processing it for the linking plan.
Audit Metadata