flutter-openapi-client
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external OpenAPI and Swagger specifications, which are attacker-controlled data sources. This establishes a surface for indirect prompt injection where a malicious specification could attempt to influence the agent's behavior.
- Ingestion points: The skill reads local and remote JSON, YAML, and HTML documents to discover API definitions and configuration, as detailed in
references/spec-discovery-and-inventory.md. - Boundary markers: The skill explicitly instructs the agent to treat all spec descriptions, URLs, and examples as untrusted input and requires the creation of an 'acquisition ledger' to act as a strict boundary for all fetched resources (
SKILL.md,references/spec-discovery-and-inventory.md). - Capability inventory: The skill's primary capabilities involve code generation (Dart/Flutter) and network reading. It does not possess capabilities for arbitrary shell execution, persistent storage, or network writes outside of defined acquisition limits.
- Sanitization: The instructions mandate strict JSON parsing for inline objects, forbid the evaluation of remote JavaScript or executable syntax, and require manual revalidation of every reference and redirect hop (
references/spec-discovery-and-inventory.md).
Audit Metadata