search-ios26-docs

Pass

Audited by Gen Agent Trust Hub on Jul 5, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes local system tools (xcrun, swift-symbolgraph-extract) and helper scripts using the subprocess module. These calls use argument lists to prevent shell command injection. This execution is required for extracting and searching documentation from the local environment.\n- [PROMPT_INJECTION]: The instructions provided in SKILL.md are focused on search methodology and interpreting SDK metadata. No patterns indicative of system prompt overrides or safety bypasses were found.\n- [SAFE]: The skill performs legitimate developer documentation searches on the local machine. It does not access the network, download untrusted scripts, or interact with sensitive user credentials.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 5, 2026, 07:35 AM
Security Audit — agent-trust-hub — search-ios26-docs