swaps-cpu-profile-audit

Pass

Audited by Gen Agent Trust Hub on Sep 11, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The scripts/run.sh script executes several system commands including unzip, tar, mv, and node to stage and process profile data. It also invokes yarn react-native-release-profiler to convert Hermes profiles. These actions are aligned with the skill's primary function of performance analysis.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external CPU profile data (JSON) provided by the user. This creates a surface where malformed or malicious profile data could influence the agent's analysis, although the processing is handled through a standalone Node.js script using standard parsing.
  • Ingestion points: User-provided .cpuprofile or -converted.json files are read by scripts/run.sh and scripts/analyze-cpuprofile.cjs.
  • Boundary markers: Not explicitly implemented in the prompt, but the data is structured and processed by an external script.
  • Capability inventory: Includes file system access (read/write within a temporary run folder) and execution of project dependencies via yarn.
  • Sanitization: Relies on standard JSON.parse() for ingesting profile data.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 11, 2026, 07:59 PM
Security Audit — agent-trust-hub — swaps-cpu-profile-audit