docs-artifact
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process information from untrusted external sources, creating a potential surface for indirect prompt injection.
- Ingestion points: The instructions describe workflows for reading data from existing documents, external Slack threads via URLs, and meeting agendas (SKILL.md).
- Boundary markers: The skill explicitly instructs the agent to check existing access permissions and follow a
<confirmation_policy>before adding sensitive content to shared documents, which serves as a procedural boundary. - Capability inventory: The skill possesses the ability to read local and remote files, create/edit documents in various formats (Word, PDF, Google Docs), and invoke shared skills for document rendering and export.
- Sanitization: There is no mention of programmatic sanitization or filtering of external content; safety relies on the mentioned confirmation policies and access checks.
Audit Metadata