doc-coauthoring
Pass
Audited by Gen Agent Trust Hub on Mar 1, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: The skill contains a surface for indirect prompt injection due to its document-fetching design.
- Ingestion points: The workflow in 'Stage 1: Context Gathering' (SKILL.md) explicitly instructs the agent to fetch content from user-provided links, shared documents, and team threads.
- Boundary markers: The instructions do not implement delimiters or 'ignore embedded instructions' warnings for the data being fetched.
- Capability inventory: The skill utilizes external URL fetching tools and interacts with the
word-documentsskill to create and edit files. - Sanitization: No sanitization or validation logic is defined to check the content of external documents before they are processed by the agent.
- [EXTERNAL_DOWNLOADS]: The skill requests the agent to perform network operations to fetch data from external URLs and shared document platforms (e.g., Google Docs, internal wikis) when provided by the user during the context gathering phase.
Audit Metadata