ash-mcp

Pass

Audited by Gen Agent Trust Hub on Nov 14, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes findings from external codebases and configuration files, which could contain malicious instructions designed to influence the agent's behavior. The results include code snippets and descriptions that are presented to the agent. Evidence includes: ingestion points via the source_dir argument, repositories cloned through set_source_git, and zip archives; capability inventory includes reading file contents and performing git operations; and the instructions lack explicit guidance on using boundary markers or sanitizing the content of findings for the agent context.\n- [COMMAND_EXECUTION]: The skill facilitates the retrieval of remote code via set_source_git to enable auditing of remote repositories. Access control is managed through a configurable whitelist of permitted git hosts.\n- [DATA_EXFILTRATION]: The skill allows for reading files and scan results from the server's local filesystem, creating a potential path for data exposure if the agent is directed to scan sensitive directories. The server implementation is documented to mitigate this by enforcing root directory restrictions and refusing access to system directories.\n- [SAFE]: The skill documentation provides installation instructions that point to the official awslabs repository. The recommended command uses uvx to fetch the tool from its official home on GitHub.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 14, 2026, 12:41 PM