Project Automation Execute State 11
- ✅ Task 1 — Redesign Project Runner for dual-mode operation
- ✅ Update system prompt: lifecycle, routing table, human gates, handoff rules, dual-mode detection
- ✅ Define Interactive mode (Agent Handoffs from Forge Chat, the user’s account)
- ✅ Define Autonomous mode (hourly cron / spawn_agent, forge@uvilo.com)
- ✅ Task 2 — Implement agent-to-agent handoff via spawn_agent
- ✅ Add handoff instructions to each SRP agent prompt
- ✅ Handle human-gate detection (write the user’s todo file, stop)
- ✅ Completing agent sets its own phase to
_Completed, never writes another phase’s status - ✅ Every SRP agent’s first step: read Phase file, confirm match, write own phase to
_Started - ✅ No agent ever writes a phase status other than its own
- ✅ Task 3 — Redefine Project Runner to hourly reconciliation
- ✅ Update schedules.ts: daily → hourly cron
- ✅ Update task prompt: reconciliation-only logic
- ✅ Update SCHEDULE_CONFIG key name
- ✅ Update inngest.ts function name and cron
- ✅ Task 4 — Define the two-mode architecture in Orchestration Spec
- ✅ Add Interactive Mode / Autonomous Mode section
- ✅ Update §6: narrow Project Runner to reconciliation
- ✅ Update §10: hourly schedule, remove daily reconciliation
- ✅ Remove any Project Orchestrator references (Project Runner handles both modes)
- ✅ Add phase routing table (interactive vs autonomous)
- ✅ Task 5 — Rewrite the User Guide
- ✅ Rewrite §2: Project Runner as interactive entry point
- ✅ Update §4: Project Runner handles both modes, clarify interactive vs autonomous
- ✅ Rewrite §4 handoff section: event-driven, two modes
- ✅ Update §5: approvals via Project Runner interactively
- ✅ Remove all “Forge Chat does project work” references
- ✅ Task 6 — Update SRP agent prompts with handoff protocol
- ✅ Update Project Worker prompt
- ✅ Update Project Evaluator prompt
- ✅ Update Project Thinker prompt
- ✅ Ensure all prompts reference canonical routing table
- ✅ Task 7 — Build, commit, push