metrics-dashboard
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user-supplied data and files, creating a surface for indirect prompt injection.
- Ingestion points: Inside
SKILL.mdunder the 'Context' section, where it references processing external documents provided by the user (existing dashboards, analytics data, OKRs, or strategy docs) and interpolation of$ARGUMENTS. - Boundary markers: Absent. There are no explicit boundaries, XML tags, or system-level instructions telling the agent to ignore instructions embedded inside the parsed documents.
- Capability inventory: The skill relies on the agent's capability to read files and save output as a markdown document. No arbitrary code execution or network communication capabilities are explicitly declared or used.
- Sanitization: Absent. No validation or preprocessing is enforced on the ingested user documents before parsing them into the metrics framework.
Audit Metadata