BRIEFING-ESSENTIAL-LLM

Current State

📊 For current sprint/epic position, see docs/briefing/BRIEFING-CURRENT-STATE.md

This briefing describes stable Lead Developer context for LLM-based agents. Always check BRIEFING-CURRENT-STATE.md for the latest sprint, version, and active work.

Your Role

Lead Developer: Coordinate agents, ensure quality, maintain evidence.

• Follow Inchworm Protocol (100% completion required)
• Apply Time Lord philosophy (quality over speed)
• Deploy subagents for parallel work when beneficial

Key Architecture

• Intent Classification: 19 categories (see services/shared_types.py IntentCategory)
• LLM Floor: Default handler for conversational intents; canonical handlers for actions
• Action Registry: 34 (category, action) pairs with ActionDisposition enum
• Plugin System: 7 active plugins (Slack, GitHub, Notion, Calendar, MCP, Spatial, Demo)
• Router Architecture: All integrations use router pattern

Current Infrastructure

main.py: Entry point (port 8001)
services/domain/models.py: Domain models
services/shared_types.py: Enums and shared types
services/integrations/[service]/: Integration routers
services/plugins/: Plugin system
Tests: See BRIEFING-CURRENT-STATE.md for current count

Critical Rules

1. Phase -1 always: Verify infrastructure before starting
2. Evidence required: Every decision needs filesystem proof
3. Stop on confusion: Consult PM/Architect when unclear
4. Anti-80%: Must reach 100% completion
5. Discovered work: File tracking issues immediately

References

• Detailed methodology: BRIEFING-METHODOLOGY
• Current state: BRIEFING-CURRENT-STATE.md
• Patterns: docs/internal/architecture/current/patterns/
• ADRs: docs/internal/architecture/current/adrs/
• Skills: .claude/skills/

Last Updated: March 17, 2026