topic-cluster-planning
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is composed of Markdown instructions and reference materials. No executable code, scripts, or suspicious command patterns are present within the skill files.
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of external data by instructing the agent to research search engine results and competitor content, which is a known surface for indirect prompt injection.
- Ingestion points:
SKILL.md(Step 1) requires the agent to read top search results, "People Also Ask" questions, and competitor publications. - Boundary markers: The instructions do not include specific delimiters or warnings to ignore instructions embedded in external content.
- Capability inventory: The skill is limited to generating text-based content plans and does not have capabilities for file modification, network exfiltration, or arbitrary command execution.
- Sanitization: No sanitization or validation logic is applied to the researched external data.
Audit Metadata