using-agent-brain
Pass
Audited by Gen Agent Trust Hub on Mar 27, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The documentation references software acquisition from well-known and trusted sources, including the Python Package Index (PyPI) and the official
uvtool installer hosted atastral.sh.- [REMOTE_CODE_EXECUTION]: The installation guide includes a standard pattern for theuvtool which involves piping a remote shell script to the local interpreter from a well-known domain (astral.sh).- [COMMAND_EXECUTION]: The skill provides commands for server lifecycle management, automated plugin installation into specific agent runtime directories, and metadata enrichment via theinjectcommand which executes local Python scripts.- [COMMAND_EXECUTION]: Troubleshooting documentation includes standard administrative commands for system package management (sudo apt-get) and process control (kill).- [PROMPT_INJECTION]: The skill implements a Retrieval-Augmented Generation (RAG) workflow that indexes and retrieves content from external documents and source code, creating a surface for indirect prompt injection. - Ingestion points: The
indexandfolders addcommands process local Markdown, PDF, and source code files. - Boundary markers: Search results are delivered to the agent in structured JSON format.
- Capability inventory: The skill is permitted to use the
BashandReadtools. - Sanitization: The documentation focuses on retrieval relevance and depth; no automated content filtering or sanitization of indexed text is specified.
Audit Metadata