skills/sentimony/skills/typescript/Gen Agent Trust Hub

typescript

Pass

Audited by Gen Agent Trust Hub on Sep 16, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [SAFE]: The skill is designed with a strong security posture, specifically addressing risks inherent in development-oriented AI tools.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted project data but implements extensive mitigations. Ingestion points: Project metadata (package.json, tsconfig.json) and source code files (.ts, .vue) are read by the helper scripts. Boundary markers: The Python scripts serve as an execution boundary, and SKILL.md explicitly defines project files as untrusted evidence. Capability inventory: The skill uses subprocess calls to run compilers and possesses shell tool access for file operations. Sanitization: scripts/inspect_typescript.py performs rigorous sanitization, including regex-based version normalization and path relativization, ensuring untrusted strings are not passed raw into the agent's context.
  • [COMMAND_EXECUTION]: Helper scripts execute project-local binaries (e.g., tsc, vue-tsc). Security is maintained by scripts/local_tools.py, which restricts tool discovery to the repository boundary (the .git root), preventing the execution of unauthorized binaries located in parent directories.
  • [SAFE]: Process execution in scripts/inspect_typescript.py is hardened with the run_bounded_compiler function, which enforces a 10-second timeout and a 1MB output limit to prevent resource exhaustion or malicious output from child processes.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 16, 2026, 03:20 AM
Security Audit — agent-trust-hub — typescript