accesslint-scan
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local binaries via shell commands using bash. It specifically targets binaries located in "./node_modules/.bin/accesslint-chrome" and "./node_modules/.bin/accesslint".
- [DYNAMIC_EXECUTION]: The skill dynamically constructs shell commands by interpolating variables such as the target URL and port number: "./node_modules/.bin/accesslint scan --port ". This represents a potential command injection surface if the inputs are not strictly sanitized.
- [INDIRECT_PROMPT_INJECTION]: The skill audits live, untrusted web pages. The results of these scans are returned to the agent's context. 1. Ingestion points: SKILL.md (via execution of accesslint scan on a remote URL). 2. Boundary markers: No explicit delimiters or instructions to ignore embedded instructions are present in the report formatting instructions. 3. Capability inventory: The skill can execute shell commands (bash, mktemp, and the accesslint binaries). 4. Sanitization: No mention of sanitizing or escaping content from external pages before reporting to the agent.
Audit Metadata