contract-consistency-check-free

Pass

Audited by Gen Agent Trust Hub on Sep 16, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill's primary function is to process user-provided contract text and return findings that include verbatim snippets from the input. This creates a risk where malicious instructions embedded in a contract (e.g., hidden in comments, placeholders, or defined terms) could be interpreted by the AI agent as valid commands when it processes the extraction results.\n
  • Ingestion points: The run.mjs script reads contract content from local files or JSON payloads provided via the --input argument.\n
  • Boundary markers: The skill does not provide explicit delimiters or instructions to the agent to treat the analyzed contract content strictly as data, increasing the risk of instruction following from untrusted input.\n
  • Capability inventory: The skill has permissions to use the Bash tool for script execution and the Read tool for accessing contract files on the filesystem.\n
  • Sanitization: The engine extracts specific entities using regular expressions but does not perform sanitization of the extracted text snippets to prevent the interpolation of malicious instructions into the agent's context.\n- [EXTERNAL_DOWNLOADS]: The skill's documentation and help output include instructions to access external URLs for tool installation and payment services.\n
  • Evidence: SKILL.md contains links to skillpay.alipay.com and a vendor resource at chenqg618.github.io for purchasing additional capabilities.\n- [REMOTE_CODE_EXECUTION]: The documentation provides specific commands for the agent to download and execute code from a remote package registry.\n
  • Evidence: SKILL.md includes a prompt instruction for the agent to run npx -y @alipay/agent-payment@latest install. This command results in the dynamic download and execution of code from the NPM registry. The package scope is associated with a well-known service provider.\n- [COMMAND_EXECUTION]: The skill requires executing local scripts and system commands to perform its analysis.\n
  • Evidence: The skill is designed to be invoked via the Bash tool to run the node scripts/run.mjs processing script.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 16, 2026, 11:29 AM
Security Audit — agent-trust-hub — contract-consistency-check-free