dart-use-path-package

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill describes how to handle untrusted path data from external sources, identifying a potential injection surface.
  • Ingestion points: Input parameters filePath, projectRoot, and filename in the Dart examples.
  • Boundary markers: None explicitly for the input data, but the skill mandates the use of package:path tools (p.split, p.join) which effectively mitigate many path-based injection attacks.
  • Capability inventory: The provided examples perform only string manipulation and memory-based file system operations; there are no network, OS command, or persistent file system write capabilities.
  • Sanitization: Lexical normalization and canonicalization are performed using p.normalize() and p.canonicalize().
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 11:07 PM
Security Audit — agent-trust-hub — dart-use-path-package