embedded-debugging
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 free-text problem reports, support tickets, and field data, which creates an indirect prompt injection surface.
- Ingestion points: Free-text input containing symptoms, logs, and ticket summaries parsed in
SKILL.md. - Boundary markers: Absent. No clear delimiters or constraints are specified to prevent the model from obeying instructions embedded within the problem reports.
- Capability inventory: The skill includes file system capabilities to create directories and write standalone HTML analysis reports under the
analysis/folder. - Sanitization: Absent. Input data is analyzed directly for classification without validation or formatting sanitization.
Audit Metadata