swift-accessibility-agent
Pass
Audited by Gen Agent Trust Hub on Mar 14, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No executable code is shipped with the skill. The repository consists entirely of Markdown documentation, JSON manifests, YAML configurations, and text-based plan files.
- [SAFE]: The skill incorporates a robust safety layer through AGENT.md and security-policy.json, which provide explicit directives to the agent to deny arbitrary code execution, exfiltration, and unauthorized system configuration changes.
- [SAFE]: All external URLs and source references point to official Apple documentation, design award pages, or well-established developer community sites (e.g., SwiftLee, Hacking with Swift, and Deque). These are treated as safe sources per trusted service recognition rules.
- [SAFE]: Indirect prompt injection risks (Category 8) are inherently addressed as the ingestion of user-provided UI code is the primary task of the skill. This surface is guarded by instructions to treat external content as untrusted and the absence of high-privilege tool capabilities within the skill's design.
Audit Metadata