Zero-config two-agent coding loop: Claude Code implements, Codex reviews, and Claude fixes until APPROVED.
Autonomous AI agent execution with iterative planning - runs agents in a continuous loop
Agentic loop for Rust — multi-turn tool dispatch, streaming, and conversation management over any LLM provider
Rust agent framework based on harness-engineering principles (Lopopolo, Böckeler). Facade re-exporting the public surface of the harness-rs-* crates.
Deterministic + agent hybrid state-machine executor for harness-rs. Stripe-Minions-style with branch-on-failure retries.
Command-line interface for harness-rs — new, skills validate/list/lint/export, trace, mcp serve.
Five-stage progressive context compactor for harness-rs — DefaultCompactor (structural) and ModelBackedCompactor (semantic).
Default World runtime (Tokio process runner, system clock, in-memory KV) and Context helpers for harness-rs.
Core traits and types for the harness-rs agent framework: Model, Tool, Guide, Sensor, Hook, Compactor, Skill, plus Context, World, Block, Event, FixPatch and 27 lifecycle events.
Optional standalone scheduler for the harness-rs agent framework. Runs declarative TOML-configured jobs (daily / weekly / interval / cron) that invoke any shell command — typically a harness-rs-built agent binary.
HookBus for harness-rs. Dispatches 27 lifecycle events with Allow/Deny/Inject/Mutate semantics. Optional OpenTelemetry feature.
ReAct agent loop, subagent isolation, and session record/replay (JSONL) for the harness-rs framework.
Procedural macros for harness-rs: #[skill], #[tool], #[guide], #[sensor], #[hook]. agentskills.io-compliant registration via inventory.
MCP (Model Context Protocol) stdio server exposing the harness-rs tool registry to Claude Code, Cursor, and Codex.