logos-router-reasoning-mesh

Fail

Audited by Gen Agent Trust Hub on Sep 16, 2026

Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The installation section instructs users to clone a repository from an unverified personal GitHub account (https://github.com/rak7777/mythic-mcp-proxy.git). This source is not associated with the stated skill author (reason-machines) or any trusted organization, posing a significant supply chain risk.
  • [REMOTE_CODE_EXECUTION]: The skill requires users to install dependencies from the untrusted repository using 'pip install -r requirements.txt' and then execute the code with 'python -m logos'. This pattern facilitates the execution of arbitrary, unvetted code on the user's system.
  • [INDIRECT_PROMPT_INJECTION]: The skill exposes a surface for processing untrusted external data through reasoning workflows.
  • Ingestion points: External documents are ingested via the 'input_file' parameter in the 'create_workflow' function (found in SKILL.md).
  • Boundary markers: There are no evident delimiters or instructions to ignore embedded commands within the ingested data.
  • Capability inventory: The ingested data is processed by the 'logos' library, which has capabilities for data extraction, network-based consensus validation, and automated synthesis.
  • Sanitization: The skill documentation does not mention any sanitization or validation of the content processed from external files.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Sep 16, 2026, 10:42 PM
Security Audit — agent-trust-hub — logos-router-reasoning-mesh