typed-structures

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill identifies a dependency on the @ohm-js/cli package and provides instructions to install it globally using pnpm if it is missing. This follows standard developer tool management practices.
  • [COMMAND_EXECUTION]: The provided Node.js script (scripts/check-syntax.mjs) executes the ohm command-line utility via spawnSync to perform syntax validation on content extracted from markdown blocks.
  • [DYNAMIC_EXECUTION]: The syntax checker script dynamically determines the binary path for the validation tool by checking the OHM_BIN environment variable at runtime.
  • [INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it is designed to ingest and review untrusted notation design drafts from users.
  • Ingestion points: Notation code snippets located in examples.md or provided by the user during a session.
  • Boundary markers: The skill uses markdown code blocks (e.g., ```spore) to delimit notation snippets from other text.
  • Capability inventory: The skill can execute local commands via the check-syntax.mjs script.
  • Sanitization: The script writes content to a temporary file before passing it to the validator, preventing direct shell interpretation of the input content.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 02:26 AM
Security Audit — agent-trust-hub — typed-structures