Per-project plan orchestration via MCP

Coding agents can now list, claim, and complete plans for one project through a limited MCP surface — without full shell, silent promote, or unbounded spend.

Anchor already had /work, headless work_once, and two MCPs that were either doctrine (anchor-prompts) or global fleet endpoints (model-fleet). What was missing was a project-bound tool plane for .plans/ itself. The new mcp/project-orchestrator/ server fills that gap.

Bind one process to one app root:

python mcp/project-orchestrator/server.py \
  --project /path/to/myapp \
  --agent-id cursor-mid-1 \
  --tier mid

Or register it with Claude Code / any stdio MCP client using the same flags. Optional project config lives at .anchor/mcp.yaml (agent_id, worker_tiers, stale_after, capabilities).

v1 is deliberately L0+L1. Agents get inventory, plan read, conventions, heuristic dependency suggestions (propose only — no LLM, no plan-file writes), stale/tier-gap warnings, claim/release, and move-only complete after the client asserts Done when. Selection and leases reuse plan_select / plan_lease — the same rules as /work. Promote stays human-only. L2 handoff, L3 orchestrate, L4 allowlisted shell, and loopback dashboard stats are deferred follow-ups.

Use a distinct --agent-id from systemd fleet_watch workers so leases do not fight. Full matrix and refuse rules: mcp/project-orchestrator/README.md. Docs overview: MCP servers.

Alongside this, agent process rules are stricter for any Anchor-scaffolded project: run /commit-prep before every commit (project-agnostic tests/changelog/blog), and open feature branches from dev/develop — creating dev from main/master when the integration branch is missing. Those are operator/agent discipline, not new product binaries, but they keep multi-agent work reviewable.