creating-sdk-apps

Pass

Audited by Gen Agent Trust Hub on Sep 4, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Fetches documentation from Anthropic's official domain and checks package metadata from well-known registries including npmjs.com and pypi.org.
  • [COMMAND_EXECUTION]: Executes standard package management and build commands including npm, pip, and npx tsc as part of the project scaffolding and verification process.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user input such as project names and agent descriptions and interpolates them into file generation and project setup logic.
  • Ingestion points: User-provided strings for project name and agent type in SKILL.md.
  • Boundary markers: None explicitly defined for the interpolated user strings.
  • Capability inventory: File system writes, package installations, and type-checking command execution.
  • Sanitization: None specified for user-provided project metadata before use in file creation.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 4, 2026, 08:33 AM
Security Audit — agent-trust-hub — creating-sdk-apps