doc-coauthoring
Pass
Audited by Gen Agent Trust Hub on Aug 20, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill acts as a surface for indirect prompt injection by ingesting untrusted data from external collaboration tools and files.\n
- Ingestion points: Team communication channels (Slack, Teams) and document storage (Google Drive, SharePoint) are specified in SKILL.md as primary sources for document context.\n
- Boundary markers: The instructions lack specific delimiters or explicit instructions for the agent to disregard commands potentially embedded in the fetched context or documents.\n
- Capability inventory: The skill has the ability to modify the filesystem using
create_fileandstr_replacetools to draft and refine document sections.\n - Sanitization: There is no evidence of filtering, escaping, or validation of the external content before it is processed by the agent or used to update files.
Audit Metadata