logos-router-reasoning-mesh
Warn
Audited by Gen Agent Trust Hub on Jul 1, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The installation guide directs users to clone a repository from an unverified GitHub account (
rak7777/mythic-mcp-proxy). This account does not align with the stated skill author (Aradotso) and is not a recognized trusted organization. - [EXTERNAL_DOWNLOADS]: The skill requires the installation of the
logos-routerpackage from a public registry. This dependency is not from a well-known or trusted vendor, increasing the risk of software supply chain interference. - [PROMPT_INJECTION]: The skill processes untrusted external data, including user prompts, PDF documents, and source code files, creating a vulnerability to indirect prompt injection.
- Ingestion points: Untrusted data enters the agent context through
router.query(prompt=...),router.create_workflow("research_paper.pdf"), andopen("auth.py").read()in SKILL.md. - Boundary markers: The instructions do not define boundary markers or safety delimiters for the processed content.
- Capability inventory: The skill uses subprocess calls via the
logos-routerCLI and network operations through REST and WebSocket APIs (http://localhost:8080,ws://localhost:8081). - Sanitization: No evidence of sanitization or validation of the external content before processing is provided in the instructions.
Audit Metadata