Expand description
Hunt Scan — MCP agent scanning and vulnerability detection.
This crate provides the core scanning engine for clawdstrike hunt scan.
It discovers AI agent MCP configurations, introspects servers, and detects
vulnerabilities in tool descriptions and configurations.
Modules§
- analysis
- Vulnerability analysis — verification API client and local heuristic checks.
- discovery
- Platform-aware AI-agent client discovery.
- mcp_
client - MCP client for server introspection.
- models
- Data models for the hunt scan subsystem.
- packages
- Direct package scanning — scan MCP servers from package registries.
- policy_
eval - Policy evaluation — check discovered MCP tools against clawdstrike guards.
- query_
filter - Natural-language keyword query filter for scan results.
- receipt
- Signed scan receipt generation using
hush-core. - redact
- Sensitive data redaction.
- skills
- Skill directory scanning — SKILL.md discovery and entity mapping.
- storage
- Persistent scan history and SHA-256 change detection.