ptlam-health-connector-data-type
Pass
Audited by Gen Agent Trust Hub on Aug 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The instructions direct the agent to execute standard development toolchains including Melos, the Dart SDK, Gradle, and Swift Package Manager for automation tasks such as code generation (Pigeon), formatting, static analysis (detekt, SwiftLint), and unit testing. These operations are performed within the project workspace following established melos scripts.
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) due to its core functionality of transforming user-provided health data type specifications into executable source code across multiple packages and platforms.
- Ingestion points: The workflow begins by resolving data type and platform contracts directly from inputs provided in the user's initial request (SKILL.md).
- Boundary markers: The skill does not define specific delimiters or instructions to ignore potential instructions embedded within the user request, increasing the risk of the agent following malicious commands included in the metadata.
- Capability inventory: The agent is empowered to modify critical source files in the facade, core, and native platform layers, and is authorized to execute build scripts and verification tasks.
- Sanitization: No explicit sanitization or strict schema validation is required for user-supplied identifiers and field names before they are interpolated into code templates.
Audit Metadata