portolan-bootstrap
Pass
Audited by Gen Agent Trust Hub on Aug 6, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes metadata from external geospatial data sources, creating a surface for indirect prompt injection. Maliciously crafted source data could influence agent behavior during the enrichment phase.\n
- Ingestion points: Remote service URLs (WFS, ArcGIS) and local geospatial files (SKILL.md).\n
- Boundary markers: The skill implements procedural checkpoints where the agent must present found data to the user for confirmation (SKILL.md).\n
- Capability inventory: The skill has the capability to execute shell commands (portolan extract) and perform network pushes (portolan push).\n
- Sanitization: The instructions focus on exact data representation and user review rather than automated sanitization.\n- [COMMAND_EXECUTION]: The skill invokes shell commands such as
portolan extract <type> "<URL>". Although the URL is quoted, the use of unvalidated parameters in shell commands presents a risk of command injection.
Audit Metadata