verify

Pass

Audited by Gen Agent Trust Hub on Aug 15, 2026

Risk Level: SAFEREMOTE_CODE_EXECUTIONCOMMAND_EXECUTION
Full Analysis
  • Remote Code Execution Pattern: The skill documentation references a common installation pattern (curl ... | bash) targeting https://claude.ai/install.sh. This is a vendor-owned domain used for distributing official CLI tools. The skill includes specific warnings against running the live installer and instead provides a sandbox harness to mock this behavior for verification purposes.
  • Command Execution (Test Harness): The skill provides a bash script intended to create a temporary test environment. This involves creating shell stubs (mocked versions of curl and bash) and executing local Python scripts from the scripts/ directory. This is standard practice for developer-oriented workflows to ensure build scripts function correctly before release.
  • Environment Isolation: The instructions use env -i and isolated HOME/PATH variables to prevent the build scripts from interacting with the user's actual system or discovering real binaries, which is a positive security practice for testing repository tools.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 15, 2026, 11:34 PM
Security Audit — agent-trust-hub — verify