doc-coauthoring
Pass
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process information from untrusted external sources, which could potentially contain malicious instructions intended to influence the agent's behavior during the drafting process.
- Ingestion points: The 'Context Gathering' stage in
SKILL.mddescribes fetching data from shared document links, messaging platforms (Slack, Teams), and uploaded files. - Boundary markers: There are no explicit instructions to use delimiters or 'ignore' directives when processing the retrieved external content.
- Capability inventory: The skill utilizes
create_fileandstr_replaceto generate and modify document artifacts or local files based on the processed context. It also has the ability to invoke sub-agents to process the drafted content. - Sanitization: There is no mention of sanitizing or validating the retrieved content for hidden malicious instructions before it is used to guide the drafting process.
Audit Metadata