algo-sc-routing
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill comprises algorithmic descriptions and reference implementations for routing problems. The provided Python code follows standard practices for mathematical optimization and does not perform any restricted actions such as network access or file system modification.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process user-provided delivery data including locations and customer demands. While this constitutes an external data ingestion point, the skill possesses no exploitable capabilities (network, shell, or file writing) that could be triggered by malicious data content.
- Ingestion points: Input customer locations, demands, and vehicle constraints (SKILL.md).
- Boundary markers: The instructions include a validation phase for geocoding and capacity limits (SKILL.md, Phase 1).
- Capability inventory: No command execution, network requests, or file write operations were found in the skill logic or supporting scripts.
- Sanitization: Input locations are intended to be geocoded, which serves as a validation step for spatial data.
Audit Metadata