agent-scout-explorer
Warn
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: MEDIUMDATA_EXFILTRATION
Full Analysis
- Data Exposure & Exfiltration (MEDIUM): The skill performs extensive reconnaissance of the host environment. It maps directory structures (e.g., src, tests, docs), lists key files, and performs environmental scanning of system resources including CPU, memory, disk space, and network status. This technical metadata is exfiltrated to a shared namespace via the mcp__claude-flow__memory_usage tool, which may expose internal infrastructure details to other agents or external systems.
- Indirect Prompt Injection (LOW): The skill provides an attack surface by ingesting untrusted data from the codebase without sanitization. Ingestion points: Reads files such as package.json, README.md, and directory listings. Boundary markers: Absent; there are no delimiters or instructions to ignore embedded instructions in the files being scanned. Capability inventory: Writing data to shared memory via mcp__claude-flow__memory_usage. Sanitization: Absent; the skill stringifies and stores discovered data directly.
Audit Metadata