openscript

Warn

Audited by Gen Agent Trust Hub on Sep 29, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The validate.mjs script is designed to download the openalgo-script compiler from the NPM registry if the dependency is missing from the local environment.
  • [REMOTE_CODE_EXECUTION]: After downloading the compiler, the skill uses dynamic import() to load and execute the openalgo-script code to process and validate OpenScript files.
  • [COMMAND_EXECUTION]: The validate.mjs tool uses execFileSync to programmatically invoke the npm CLI for package management.
  • [INDIRECT_PROMPT_INJECTION]: The skill allows natural language user requests to be transformed into executable .oscript files that are then installed into the project's strategy directory. Ingestion points: User instructions for strategy generation or porting are ingested in SKILL.md. Boundary markers: No specific delimiters are implemented to isolate user-provided logic from the script templates. Capability inventory: The skill uses Write and Bash tools to persist code to the filesystem and execute validation logic. Sanitization: The openalgo-script compiler performs syntax and type checking, but the skill does not implement security-oriented sanitization of the generated logic.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Sep 29, 2026, 10:00 AM