collaborative-document-creation
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process content from external sources such as shared documents, Slack/Teams messages, and cloud storage files. It lacks explicit boundary markers or sanitization instructions for this untrusted data, which could allow instructions embedded in those sources to influence the agent.
- Ingestion points: SKILL.md (Stage 1) mentions fetching content from user-provided links, files, and integrations with messaging and storage apps.
- Boundary markers: The instructions do not define specific delimiters or "ignore instructions" wrappers for the external content.
- Capability inventory: The skill uses
create_fileandstr_replacefor document assembly, invokes sub-agents for testing, and leverages integrations to read organizational data. - Sanitization: No explicit validation or filtering of the fetched external content is described.
Audit Metadata