port-blueprints
Pass
Audited by Gen Agent Trust Hub on Jul 15, 2026
Risk Level: SAFENO_CODEPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is authored by port-labs and provides documentation and JSON templates for modeling Port blueprints. All external URLs point to official Port domains (port.io), which are recognized as well-known service resources.
- [NO_CODE]: The skill does not contain any executable scripts or binaries, consisting only of Markdown documentation and JSON data templates, which significantly reduces the potential attack surface.
- [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection as it facilitates a workflow that involves reading existing blueprint data from the Port API and then performing follow-up actions like updates or upserts.
- Ingestion points: Reading blueprint data from GET /v1/blueprints/ or the list_blueprints MCP tool.
- Boundary markers: None identified; the agent is not instructed to use specific delimiters when processing retrieved blueprint content.
- Capability inventory: The skill provides instructions and tools for modifying or deleting Port blueprints (upsert_blueprint, curl operations).
- Sanitization: None identified; there are no instructions for validating or sanitizing retrieved blueprint definitions before the agent incorporates them into its tasks.
Audit Metadata