brainstorming-with-explorer-challenger-teams
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it incorporates external, untrusted data into the prompts of spawned subagents.
- Ingestion points: Untrusted data is gathered from the project's README, roadmap, planning files, changelog, and open issues as described in Step 1 of the SKILL.md file.
- Boundary markers: The instructions lack specific delimiters or warnings for the subagents to ignore potential instructions embedded within the ingested context.
- Capability inventory: The skill uses Teammate and Task tools to create subagents, performs codebase exploration, and writes refined reports to the filesystem.
- Sanitization: The ingested data is condensed into a context block and interpolated directly into prompt templates without visible sanitization or validation logic.
Audit Metadata