accessibility-inspect

Pass

Audited by Gen Agent Trust Hub on Aug 3, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructions advise users to install the chrome-devtools-mcp and @accesslint/cli packages via npx. These are official vendor resources or standard platform tools required for the skill's operation.
  • [COMMAND_EXECUTION]: The documentation includes shell command examples for environment setup, such as adding MCP tools to the agent and initializing project configuration files like accesslint.config.json.
  • [PROMPT_INJECTION]: As a web-browsing skill, it inherently possesses a surface for indirect prompt injection by processing external page content.
  • Ingestion points: The agent reads the DOM and accessibility tree via take_snapshot and evaluate_script calls.
  • Boundary markers: The skill does not define specific delimiters or instructions to ignore embedded commands within ingested web content.
  • Capability inventory: The skill uses chrome-devtools for browser interaction and has access to the Bash tool.
  • Sanitization: No content sanitization is performed on the ingested web data before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 3, 2026, 03:51 PM
Security Audit — agent-trust-hub — accessibility-inspect