draft
Pass
Audited by Gen Agent Trust Hub on Aug 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external message threads and user-provided briefs, which are potential vectors for indirect prompt injection attacks.
- Ingestion points: The skill ingests data from a "brief" or "a thread" to generate ready-to-send communication (SKILL.md).
- Boundary markers: There are no explicit instructions for the agent to use delimiters or specific safety markers to differentiate between user instructions and untrusted data in threads.
- Capability inventory: The skill is authorized to read voice profiles (
references/voice.md) and project documentation (context/,wiki/), use a graph query tool (/graph-query), and autonomously send internal team messages (SKILL.md). - Sanitization: The skill mandates a "Verification Gate" to ensure the accuracy of facts and figures, but it does not define protocols for sanitizing or ignoring potential instructions embedded within the message threads it processes.
Audit Metadata