IntelAgent

Standalone Rust workspace for agent orchestration and worker boundaries.

The current verified scope is the foundation layer:

• intelagent-core defines Agent, Task, Context, QualityGate, and proof types.
• PhantomWorker provides the first real Layer 0 boundary by calling Phantom over HTTP.
• The remaining crates are still placeholder modules and should be treated as scaffolding.

Current Architecture

intelagent/
├── Cargo.toml
├── README.md
├── MANIFEST.md
├── PROJECT_STRUCTURE.md
├── crates/
│   ├── cli/
│   ├── core/
│   │   ├── src/agent.rs
│   │   ├── src/context.rs
│   │   ├── src/lib.rs
│   │   ├── src/phantom_worker.rs
│   │   ├── src/proof.rs
│   │   ├── src/quality.rs
│   │   ├── src/task.rs
│   │   └── src/types.rs
│   ├── governance/
│   ├── mcp/
│   ├── memory/
│   ├── quality/
│   └── security/
└── examples/
    └── basic_agent.rs

Verified Boundary with Phantom

PhantomWorker lives in crates/core/src/phantom_worker.rs and targets Phantom's REST API:

• GET /health
• POST /api/chat

This keeps IntelAgent standalone while allowing Phantom to act as a worker backend instead of a linked library dependency.

Build and Test

This repository does not currently carry its own flake.nix. In the umbrella workspace, the validated path is:

nix develop --command bash -lc 'cd /home/kernelcore/master/intelagent && cargo test -p intelagent-core'

You can also use a compatible Rust toolchain directly:

cd /home/kernelcore/master/intelagent
cargo check
cargo test -p intelagent-core

Status

Area	Status	Notes
crates/core	functional foundation	Core abstractions plus PhantomWorker REST boundary
crates/core/tests	real integration coverage	Mocked end-to-end HTTP test for PhantomWorker
crates/cli	scaffold	Module exists, no CLI workflow yet
crates/governance	scaffold	dao.rs and rewards.rs are placeholders
crates/mcp	scaffold	Placeholder library only
crates/memory	scaffold	Placeholder library only
crates/quality	scaffold	Placeholder library only
crates/security	scaffold	audit.rs and privacy.rs are placeholders

Near-Term Direction

• Keep the focus on worker orchestration and boundary correctness.
• Do not treat DAO, token, or ZK ambitions as release blockers for the current MVP.
• Add new workers behind explicit interfaces, following the PhantomWorker pattern.