OTP-native task-execution engine an AI orchestrator drives end to end.
Harness pulls tasks from an rmap roadmap, dispatches each to a headless coding agent (Claude Code, Cursor, Codex, Grok, Antigravity) running in an isolated git worktree, runs the target project's own check stack against the result, and reports a verified outcome back over an agent-shaped surface (MCP tools + JSON CLI).
The primary user is an AI orchestrator, not a human. The verification stack — not the agent's self-report — is the source of truth for success/failure. Every adapter is held to the same AgentAdapter behaviour and a reusable conformance suite.
Early development. The core loop is taking shape — rmap task ingestion, the
per-run worktree lifecycle, the verification runner, the AgentAdapter
behaviour with its reusable conformance suite, and five concrete adapters
(Claude Code, Codex, Cursor, Grok, and Antigravity, with a generic
timeout-enforcing run driver) are in place.
See ROADMAP.md for the full plan and current task status
(rendered from roadmap/tasks.toml by rmap).
# First time
mix deps.get
mix compile
# Typical loop
mix test
mix credo --strict
mix dialyzer
mix doctor
mix sobelow --exit --skip
# With AI-friendly output
mix test.json
mix dialyzer.json
# Dev MCP surface (Tidewave on port 4016)
iex -S mix tidewaveAll tooling is wired per the global Elixir setup conventions (Styler first, Reach for OTP analysis, etc.).
MIT (or your preferred license).