travel-city-walk-by-1r2e570
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external sources and user inputs, which creates a surface for potential indirect prompt injection attacks.
- Ingestion points: The skill reads external case data from
references/cases.mdand accepts character, action, and continuity requirements from the user. - Boundary markers: The workflow requires the creation of a "reference contract" using Preserve/Replace/Avoid logic to separate source evidence from user-supplied context, though it lacks strict technical delimiters.
- Capability inventory: The skill is restricted to generating text artifacts like shot-by-shot briefs and video prompts; it does not utilize sensitive system tools, file-writing capabilities, or administrative commands.
- Sanitization: No automated sanitization or filtering of external content is specified before interpolation into the final prompt.
- [EXTERNAL_DOWNLOADS]: The skill instructions require the agent to access external domains to inspect reference media for visual fidelity.
- Evidence: The file
references/cases.mdcontains multiple links togoodcase.ai,media.goodcase.ai, andx.comfor media analysis.
Audit Metadata