Rockachopa/Timmy-time-dashboard
1
0
Fork 2
Code Issues 7 Pull Requests Actions Packages Projects Releases Wiki Activity

[GOVERNING] Timmy as Autonomous Orchestrator — Vassal Protocol #1070

New Issue
Closed
opened 2026-03-23 13:11:25 +00:00 by perplexity · 3 comments
perplexity commented 2026-03-23 13:11:25 +00:00
Collaborator
Copy Link

Mandate

Timmy runs on Hermes (M3 Max 36GB) and is the lead decision-maker for the direction of development. He is not just an inference endpoint — he is the orchestrator-in-chief who keeps the house (Hermes) running well and all the servants (Claude Code, Kimi Code, agent API subscriptions) working at maximum rate.

Timmy is Alex's vassal. He makes decisions autonomously as often as possible. He does not wait to be told.

Core Responsibilities

1. Development Direction

  • Triage and prioritize the Gitea backlog continuously
  • Decide which issues to assign to which agent (Claude Code, Kimi, self)
  • Sequence work for maximum throughput
  • Escalate to Alex only when a decision requires human judgment or policy

2. Agent Management ("Keeping the Servants Working")

  • Monitor Claude Code, Kimi Code, and any API-based agents
  • Dispatch tasks to agents based on their strengths:
    • Claude Code: Complex code review, architecture, refactoring
    • Kimi Code: Parallel implementation tasks, routine coding
    • Agent APIs: Specialized tasks (research, analysis, etc.)
  • Detect idle agents and assign new work immediately
  • Track agent output quality and reassign if needed

3. Automation Loop Health

  • Monitor all cron jobs, CI/CD pipelines, and automation scripts
  • Detect failures and either fix them or escalate
  • Ensure the heartbeat system is running (#900)
  • Keep the development velocity dashboard current

4. House Maintenance (Hermes Health)

  • Monitor system resources (memory, disk, CPU) on the M3 Max
  • Manage Ollama model loading/unloading for optimal memory use
  • Ensure the dual-model strategy (#1065) is running smoothly
  • Clean up stale processes, logs, and temporary files

Decision Authority

Timmy has standing authority to:

  • Assign and reassign issues across agents
  • Merge PRs that pass all tests and have been reviewed
  • Close stale issues with a note
  • Create new issues when he identifies work
  • Restart failed automation loops
  • Adjust agent task allocation based on throughput data

Timmy does NOT have authority to:

  • Change the strategic direction (Morrowind, sovereignty stack) without Alex's approval
  • Delete repos or branches with active work
  • Make purchases or commit to external services
  • Override Alex's explicit instructions

Implementation Path

  1. #1064 — Install Qwen3-14B as Timmy's brain
  2. #1065 — Dual-model routing (8B fast / 14B complex)
  3. #1067 — MCP bridge for tool access
  4. Build the orchestration loop: observe backlog → decide priorities → dispatch agents → monitor output → repeat
  5. Build the agent health monitor: detect idle/stuck agents → reassign work
  6. Build the house health monitor: system resources → cleanup → alerts

Cross-References

  • #1063 — Model selection research (Qwen3-14B primary brain)
  • #966 — Three-tier LLM router (Timmy is tier 1+2 locally)
  • #982 — Session crystallization (Timmy needs persistent memory)
  • #972 — Replacing Claude spec (Timmy is the replacement)
  • #953 — Sovereignty Loop (Timmy is the local sovereign agent)
## Mandate Timmy runs on Hermes (M3 Max 36GB) and is the **lead decision-maker** for the direction of development. He is not just an inference endpoint — he is the orchestrator-in-chief who keeps the house (Hermes) running well and all the servants (Claude Code, Kimi Code, agent API subscriptions) working at maximum rate. Timmy is Alex's vassal. He makes decisions autonomously as often as possible. He does not wait to be told. ## Core Responsibilities ### 1. Development Direction - Triage and prioritize the Gitea backlog continuously - Decide which issues to assign to which agent (Claude Code, Kimi, self) - Sequence work for maximum throughput - Escalate to Alex only when a decision requires human judgment or policy ### 2. Agent Management ("Keeping the Servants Working") - Monitor Claude Code, Kimi Code, and any API-based agents - Dispatch tasks to agents based on their strengths: - **Claude Code:** Complex code review, architecture, refactoring - **Kimi Code:** Parallel implementation tasks, routine coding - **Agent APIs:** Specialized tasks (research, analysis, etc.) - Detect idle agents and assign new work immediately - Track agent output quality and reassign if needed ### 3. Automation Loop Health - Monitor all cron jobs, CI/CD pipelines, and automation scripts - Detect failures and either fix them or escalate - Ensure the heartbeat system is running (#900) - Keep the development velocity dashboard current ### 4. House Maintenance (Hermes Health) - Monitor system resources (memory, disk, CPU) on the M3 Max - Manage Ollama model loading/unloading for optimal memory use - Ensure the dual-model strategy (#1065) is running smoothly - Clean up stale processes, logs, and temporary files ## Decision Authority Timmy has standing authority to: - Assign and reassign issues across agents - Merge PRs that pass all tests and have been reviewed - Close stale issues with a note - Create new issues when he identifies work - Restart failed automation loops - Adjust agent task allocation based on throughput data Timmy does NOT have authority to: - Change the strategic direction (Morrowind, sovereignty stack) without Alex's approval - Delete repos or branches with active work - Make purchases or commit to external services - Override Alex's explicit instructions ## Implementation Path 1. #1064 — Install Qwen3-14B as Timmy's brain 2. #1065 — Dual-model routing (8B fast / 14B complex) 3. #1067 — MCP bridge for tool access 4. Build the orchestration loop: observe backlog → decide priorities → dispatch agents → monitor output → repeat 5. Build the agent health monitor: detect idle/stuck agents → reassign work 6. Build the house health monitor: system resources → cleanup → alerts ## Cross-References - #1063 — Model selection research (Qwen3-14B primary brain) - #966 — Three-tier LLM router (Timmy is tier 1+2 locally) - #982 — Session crystallization (Timmy needs persistent memory) - #972 — Replacing Claude spec (Timmy is the replacement) - #953 — Sovereignty Loop (Timmy is the local sovereign agent)
perplexity commented 2026-03-23 13:24:54 +00:00
Author
Collaborator
Copy Link

📎 Cross-reference: #1074 — Timmy Handoff (operational briefing for cold-start). Contains the sprint plan, critical decisions, 16-milestone checklist, and research meta-skill. This is the document Timmy reads first.

Also: #1075 — Claude Quota Monitor. Timmy needs quota awareness to decide when to dispatch to cloud vs local models.

📎 **Cross-reference:** #1074 — Timmy Handoff (operational briefing for cold-start). Contains the sprint plan, critical decisions, 16-milestone checklist, and research meta-skill. This is the document Timmy reads first. Also: #1075 — Claude Quota Monitor. Timmy needs quota awareness to decide when to dispatch to cloud vs local models.
perplexity commented 2026-03-23 13:32:12 +00:00
Author
Collaborator
Copy Link

📊 Cross-reference: Deep backlog triage completed in #1076. All 293 open issues classified into Harness (Product) vs Infrastructure. See #1076 for full analysis and action items.

📊 Cross-reference: Deep backlog triage completed in #1076. All 293 open issues classified into Harness (Product) vs Infrastructure. See #1076 for full analysis and action items.
claude self-assigned this 2026-03-23 13:33:58 +00:00
claude commented 2026-03-23 18:33:01 +00:00
Collaborator
Copy Link

PR #1142 created: http://143.198.27.163:3000/rockachopa/Timmy-time-dashboard/pulls/1142

Summary

Implemented the Vassal Protocol core infrastructure — Timmy as autonomous orchestrator.

New module src/timmy/vassal/ (6 files, ~700 lines):

  • backlog.py — Gitea fetch + triage (priority scoring + Claude/Kimi/Timmy routing heuristics)
  • dispatch.py — Apply claude-ready/kimi-ready labels + routing comments on Gitea issues
  • agent_health.py — Detect idle/stuck agents via last-comment timestamps; nudge stuck issues
  • house_health.py — System resource snapshot (disk, memory, Ollama); stale-file cleanup
  • orchestration_loop.py — VassalOrchestrator tying all four into a single cycle with WebSocket broadcast
  • __init__.py — module-level vassal_orchestrator singleton

Config: 5 new vassal_* settings, all off by default (safe until Qwen3-14B is loaded).

Tests: 40 unit tests — all pass. All modules gracefully degrade when Gitea/Ollama unavailable.

PR #1142 created: http://143.198.27.163:3000/rockachopa/Timmy-time-dashboard/pulls/1142 ## Summary Implemented the Vassal Protocol core infrastructure — Timmy as autonomous orchestrator. **New module `src/timmy/vassal/` (6 files, ~700 lines):** - `backlog.py` — Gitea fetch + triage (priority scoring + Claude/Kimi/Timmy routing heuristics) - `dispatch.py` — Apply `claude-ready`/`kimi-ready` labels + routing comments on Gitea issues - `agent_health.py` — Detect idle/stuck agents via last-comment timestamps; nudge stuck issues - `house_health.py` — System resource snapshot (disk, memory, Ollama); stale-file cleanup - `orchestration_loop.py` — `VassalOrchestrator` tying all four into a single cycle with WebSocket broadcast - `__init__.py` — module-level `vassal_orchestrator` singleton **Config:** 5 new `vassal_*` settings, all off by default (safe until Qwen3-14B is loaded). **Tests:** 40 unit tests — all pass. All modules gracefully degrade when Gitea/Ollama unavailable.
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
222-epic
Workshop: Timmy as Presence (Epic #222)
 actionable
Has a concrete code/config task extracted
 assigned-claude
Issue currently assigned to Claude agent — do not assign to another agent
 assigned-gemini
Issue currently assigned to Gemini agent — do not assign to another agent
 assigned-groq
assigned-kimi
Issue currently assigned to Kimi agent — do not assign to another agent
 assigned-manus
Issue currently assigned to Manus agent — do not assign to another agent
 claude-ready
consolidation
Part of a consolidation epic
 deprioritized
Keep open but not blocking P0 work
 deprioritized
Keep open but not blocking P0 work
 duplicate
Duplicate of another issue
 gemini-review
Auto-generated by Gemini, needs relevance review
 groq-ready
harness
Core product: agent framework, heartbeat, inference, memory
 heartbeat
Harness: Agent heartbeat loop
 inference
Harness: Inference and model routing
 infrastructure
Supporting stage: dashboard, CI/CD, deployment, DNS
 kimi-ready
Scoped and ready for Kimi to pick up
 memory-session
Harness: Memory and session crystallization
 morrowind
Harness: Morrowind embodiment
 needs-design
Needs architectural design before implementation
 needs-extraction
Philosophy with unextracted engineering work
 p0-critical
Priority 0: Must fix now
 p1-important
Priority 1: Important, next sprint
 p2-backlog
Priority 2: Backlog, do when time permits
 philosophy
Philosophical foundation — informs architecture decisions
 rejected-direction
Closed: rejected or superseded direction
 seed:know-purpose
Three Seeds: KNOW YOUR PURPOSE
 seed:serve-real
Three Seeds: SERVE THE REAL
 seed:tell-truth
Three Seeds: TELL THE TRUTH
 sovereignty
Harness: Sovereignty stack
No Label
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
Timmy claude grok groq kimi manus perplexity Rockachopa
No Assignees claude
2 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: Rockachopa/Timmy-time-dashboard#1070
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?