dart-modern-features
Pass
Audited by Gen Agent Trust Hub on Sep 24, 2026
Risk Level: SAFENO_CODEINDIRECT_PROMPT_INJECTION
Full Analysis
- [NO_CODE]: The skill consists entirely of Markdown documentation and JSON evaluation configurations. It does not include any executable scripts (Python, JavaScript, shell) or binaries, which eliminates many common attack vectors like command execution or persistent backdoors.\n- [INDIRECT_PROMPT_INJECTION]: The skill provides instructions for the agent to process and refactor external Dart source code, creating a surface for indirect prompt injection via malicious code comments or data structures in analyzed files.\n
- Ingestion points: External Dart source files (e.g., network.dart, user_parser.dart) provided by the user for refactoring.\n
- Boundary markers: The skill includes 'Abstention Guardrails' and 'Pragmatic Balance' sections that explicitly define conditions under which the agent should decline or modify its refactoring behavior.\n
- Capability inventory: The agent is expected to perform file reading and modification using platform tools to implement the suggested refactors.\n
- Sanitization: The skill relies on logical constraints and abstention rules rather than explicit input sanitization of the code being processed.\n- [SAFE]: The skill includes a link to a related repository ('github.com/kevmoo/dash_skills') which is owned by the skill's author and contains further documentation. This reference is informational and does not trigger any risky network operations or remote code execution.
Audit Metadata