survey-patterns
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external sources named in the task (e.g., other repositories, branches, or stashes). This constitutes an attack surface for indirect prompt injection if the external content contains adversarial instructions intended to influence the agent's behavior.
- Ingestion points: Task descriptions and external code sources (repositories, branches, stashes) named by the user.
- Boundary markers: While the skill lacks explicit delimiters for untrusted content, it enforces a strict markdown output schema and instructions to report findings verbatim.
- Capability inventory: The subagent is restricted to read-only tools (
Glob,Grep,Read,git show) and is explicitly prohibited from writing files or proposing implementation plans in its configuration. - Sanitization: There is no explicit sanitization of the external content before it is read and summarized.
Audit Metadata