generate-native-extension

Pass

Audited by Gen Agent Trust Hub on Aug 25, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • Indirect Prompt Injection Surface: The skill ingests data from an external file, PRD.md, to drive the generation of native source code across several files (Step 1.6 and Step 3). This ingestion process represents a potential surface for indirect prompt injection, as instructions embedded in the requirement document could influence the generated logic.
  • Ingestion points: PRD.md file in the working directory.
  • Boundary markers: No explicit delimiters or warnings to ignore embedded instructions are specified for the PRD.md content.
  • Capability inventory: The skill possesses Write capabilities to create multiple native source files and Bash capabilities for structural verification.
  • Sanitization: No explicit validation or filtering of the input document's content is described.
  • External Dependency Resolution: The skill utilizes pnpm install during its smoke check phase to resolve development dependencies such as react and react-native (Step 4). This is a standard developer practice targeting the public npm registry. While this involves downloading third-party packages, it is conducted as part of the expected environment setup and does not involve the execution of unknown remote scripts.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 25, 2026, 05:29 AM
Security Audit — agent-trust-hub — generate-native-extension