How 0sec-labs/0 works
0 is an AI security harness for local and managed security reviews. A user starts a browser chat, CLI workflow, or agent connection; the engine investigates a target, produces evidence-backed findings, and supports review and fixes. Its major branches include scoped agent tools, source and binary analysis, workflow execution, findings and run storage, and integrations. The README documents product-level flows; source-index references establish selected internal dependencies. Many subsystem paths are present in the tree but their implementation and wiring were not sampled, so those relationships are omitted.
Main entry points: turn-engine.ts Chat turn engine · index.ts CLI commands · workflow-engine-service.ts Workflow runner · pipeline.py Binary analysis
Components 22
User interfaces
- CLI commands
packages/cli/src/commands/index.ts - Browser dashboard
packages/dashboard/src/main.tsx
Review engine
- Workflow runner
packages/cli/src/workflow-engine-service.ts - Chat turn engine
packages/core/src/console/turn-engine.ts - LLM runtime
packages/core/src/runtime/llm-api.ts - Agent features
packages/core/src/agent/features.ts - Scope policy
packages/core/src/scope/scope.ts
Analysis and research
- Security tools
packages/core/src/agent/tools/security-engines.ts - Active tools
packages/core/src/agent/tools/offensive-engines.ts - Unified review
packages/core/src/unified-pipeline.ts - Binary ingest
0verse/src/zeroverse/ingest.py - Binary analysis
0verse/src/zeroverse/pipeline.py - Fuzzing
0verse/src/zeroverse/fuzz/aflpp.py - CVE knowledge base
0verse/src/zeroverse/cve_kb/ingest.py
Findings and state
- Findings review
packages/cli/src/web/findings-read-api.ts - Run and finding store
packages/db/src/database.ts - Learning and memory
packages/cli/src/web/learning.ts
Agents and integrations
- Plugins and harnesses
packages/core/src/plugins/live-harness.ts - MCP connection
packages/cli/src/commands/mcp-server.ts
Other components
- Operator
- External coding agent
- Model provider
Connections 21
- Operator → Browser dashboard: starts browser chat
README.md - Operator → CLI commands: runs workflows
README.md - Browser dashboard → Chat turn engine: submits chat
- CLI commands → Workflow runner: invokes
packages/cli/src/workflow-engine-service.ts - Workflow runner → Unified review: runs review
- Chat turn engine → LLM runtime: creates runtime
packages/core/src/console/turn-engine.ts - Chat turn engine → Agent features: uses flags
packages/core/src/console/turn-engine.ts - Chat turn engine → Plugins and harnesses: hosts plugins
packages/core/src/console/turn-engine.ts - Security tools → Scope policy: checks scope
packages/core/src/agent/tools/security-engines.ts - Active tools → Scope policy: resolves scoped paths
packages/core/src/agent/tools/offensive-engines.ts - CLI commands → Findings review: reviews findings
README.md - Workflow runner → Run and finding store: persists runs
- Findings review → Run and finding store: reads findings
- Operator → Learning and memory: reviews lessons
README.md - Operator → MCP connection: connects agent
README.md - External coding agent → MCP connection: uses workflows
README.md - LLM runtime → Model provider: uses model
README.md - Binary ingest → Binary analysis: feeds analysis
0verse/src/zeroverse/pipeline.py - Binary analysis → Fuzzing: runs fuzzing
0verse/src/zeroverse/pipeline.py - Binary analysis → CVE knowledge base: uses vulnerability data
- Unified review → Findings review: returns findings