doc-coauthoring
Pass
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest large amounts of context from external sources to guide document creation, which introduces a surface for instructions embedded in that data to influence agent behavior.
- Ingestion points: As described in the 'Stage 1: Context Gathering' section of
SKILL.md, the agent ingests data from Slack, Microsoft Teams, Google Drive, SharePoint, MCP servers, and user-provided files or shared links. - Boundary markers: The instructions do not define explicit delimiters or 'ignore' instructions for the agent when processing content retrieved through these integrations.
- Capability inventory: The skill utilizes
create_fileandstr_replacetools to manage document artifacts and local files throughout the refinement stage. - Sanitization: There are no requirements for the agent to sanitize, escape, or filter instructions that might be present in the external documents or messaging threads before processing them.
Audit Metadata