icon-composer-app-icon-workflow

Pass

Audited by Gen Agent Trust Hub on Jul 31, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill invokes the ictool binary located within the Xcode application bundle and executes local validation scripts using uv and bash from the vendor's repository structure.
  • [PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection as it ingests and processes external design assets.
  • Ingestion points: Processes .icon documents, SVG/PNG source artwork, and user-provided design briefs as defined in the skill instructions.
  • Boundary markers: No delimiters or instructions to ignore embedded content are provided for the ingested assets.
  • Capability inventory: Includes execution of shell commands (ictool, bash, uv run), file system access, and GUI automation via the Computer Use capability.
  • Sanitization: The skill does not specify any sanitization or validation routines for the content of ingested files before they influence agent actions.
  • [EXTERNAL_DOWNLOADS]: References and directs the agent to fetch current guidelines and documentation from official Apple developer domains.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 31, 2026, 08:42 AM
Security Audit — agent-trust-hub — icon-composer-app-icon-workflow