roi-report
Pass
Audited by Gen Agent Trust Hub on Nov 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources, which constitutes an attack surface for indirect prompt injection if the source files are manipulated to include hidden instructions.
- Ingestion points: The skill reads telemetry data from
context/roi/audit-receipts.jsonland agent baselines fromtools/hooks/baselines.json. - Boundary markers: The instructions do not specify the use of delimiters (like XML tags or clear separators) or instructions to ignore embedded commands when parsing the external JSON/JSONL content.
- Capability inventory: The skill's impact is mitigated by its limited toolset, which only includes
ReadandGlobpermissions, preventing network exfiltration or command execution. - Sanitization: There is no evidence of data validation or sanitization before the model processes the file contents to generate the report.
Audit Metadata