neo-team
Pass
Audited by Gen Agent Trust Hub on Mar 12, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to Indirect Prompt Injection (Category 8) because it treats project-level documentation as an authoritative source of instructions for high-privilege agents.\n
- Ingestion points: The orchestrator in
SKILL.mdreads project context fromCLAUDE.md,AGENTS.md, andCONTRIBUTING.mdto define behavior for the specialist agents.\n - Boundary markers: Although the orchestration prompt uses headers (e.g.,
## Project Conventions) and delimiters (---), it lacks explicit instructions for sub-agents to ignore potentially malicious or overriding directives embedded within these files.\n - Capability inventory: The system delegates tasks to agents with
Bash,Edit, andWritecapabilities (seereferences/developer.mdandreferences/devops.md), which could be abused if the agents are successfully injected by malicious content in project files.\n - Sanitization: The skill performs no sanitization or filtering of the content extracted from the project documentation files before interpolating it into the sub-agent system prompts.
Audit Metadata