juicebox-core-workflow-b
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to ingest and analyze external candidate records and datasets from the Juicebox platform, creating a potential surface for indirect prompt injection if the data contains hidden instructions.
- Ingestion points:
SKILL.mddefines steps to query external datasets (e.g.,candidates,candidates_q1_2026) and process the resulting records. - Boundary markers: The instructions do not define delimiters or specific warnings to ignore instructions embedded within the processed candidate data.
- Capability inventory: The agent has access to
Write,Edit, andBash(npm:*)tools, which could be misused if the agent is influenced by malicious data content. - Sanitization: There are no instructions for sanitizing or validating the external content before processing or exporting it.
Audit Metadata