ops-cost-optimization
Pass
Audited by Gen Agent Trust Hub on Aug 12, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill performs legitimate infrastructure and SaaS cost auditing. All external references point to official documentation or well-known service pricing pages. The feedback loop mechanism uses a local script for internal state management. No obfuscation, data exfiltration, or unauthorized command execution was detected.
- [INDIRECT_PROMPT_INJECTION]: The skill includes an agent (agents/cost-auditor.md) that ingests untrusted external data such as billing screenshots and API responses. While this creates an attack surface for indirect prompt injection, it is the intended primary purpose of the skill, and the risk is considered low and standard for this use case.
- Ingestion points: agents/cost-auditor.md (parsing billing screenshots and usage data).
- Boundary markers: None explicitly defined, but instructions mandate verifying data against official sources in data/sources.json.
- Capability inventory: Execution of the local script append_learning.py and reading of local documentation files.
- Sanitization: Not explicitly detailed in the prompt logic.
Audit Metadata