Memory visualization
Every non-private page, markdown file and directory in the vault, plus the agent memory stores and the kanban work that touches them — drawn as one graph, the way Obsidian's Graph View draws a vault. Below it, the Agent OS panel shows what the agents working on this lab are doing.
Memory visualization
Second Brain + agent memory as one map: pages, markdown files, directories, agents and tasks, linked by contain / related / mirror / action.
loading snapshot…
Forces & local graph
Loading graph…
- page
- file
- directory
- agent
- task
- contains
- related
- mirror
- action
Text version of the graph
Agent OS
Task progress, agent states and recent actions from the kanban board that builds this lab. Read from the same build-time snapshot as the graph above — this panel never polls, so nothing here is claimed to be live.
snapshot as of
- tasks done
- 21/24
- in flight
- 1
- blocked / failed
- 0
- agents running
- 1/6
- snapshot
- 11:49Z
Agents
all agents · 44 rows
Click an agent to filter tasks and the activity log; click it again (or press Esc) to clear.
In flight 1
- t_b2e01875 ● running
Apply improvements to lab-update workflow skills
Queued 2
- t_bdc7c653 ○ todo
Send Discord closeout notification with skills, notes, mirrors
- t_8c890d38 ○ todo
close out properly.
Finished 21
- t_7e31902e ✓ done
Update profiles to enforce Discord completion notifications
- t_ed7b51c2 ✓ done
Audit lab-update workflow skills for needed updates
subtasks 0/1
- ● running Apply improvements to lab-update workflow skills
- t_3b01241d ✓ done
Write closeout notes and mirror to second brain
- t_b8df8579 ✓ done
Deploy approved /lab + /lab/hermes build (7202f79) to hetsulab.tech
- t_656a9ef7 ✓ done
Verify /lab and /lab/hermes changes against brief
subtasks 1/1
- ✓ done Deploy approved /lab + /lab/hermes build (7202f79) to hetsulab.tech
- t_5f6cb1bc ✓ done
Fix 9 pre-existing astro check type errors in AgentOsPanel.astro
- t_233253d2 ✓ done
Redesign /lab cards and add Memory/Hermes preview cards
- t_fa737422 ✓ done
Update /lab/hermes skills list
subtasks 1/1
- ✓ done Fix 9 pre-existing astro check type errors in AgentOsPanel.astro
- t_9a3bb76a ✓ done
Audit /lab and /lab/hermes structure, assets, and skills
subtasks 2/2
- ✓ done Redesign /lab cards and add Memory/Hermes preview cards
- ✓ done Update /lab/hermes skills list
- t_15a22095 ✓ done
Re-gate the corrected local-llm-hermes-desktop-rx7600 document (brief-conformance fixes)
- t_9a4fdba0 ✓ done
Review the research document against the brief
- t_22c8848f ✓ done
Reviewer round 3: gate local-llm-hermes-desktop-rx7600 rev 3 (sha256 93a16cbb…, 60,154 B)
- t_8ee85def ✓ done
Write the local-LLM-for-Hermes-Desktop research document
subtasks 1/1
- ✓ done Review the research document against the brief
- t_f3b05032 ✓ done
Writer: apply round-3 should-fix to local-llm-hermes-desktop-rx7600 rev 3 (timing-block model tag on the 9B path) → rev 4
subtasks 1/1
- ✓ done Reviewer round 3: gate local-llm-hermes-desktop-rx7600 rev 3 (sha256 93a16cbb…, 60,154 B)
- t_fa477053 ✓ done
Reviewer round 4: gate local-llm-hermes-desktop-rx7600 rev 4 (sha256 4ca4dc8e…, 60,970 B)
- t_4a8ce324 ✓ done
Reviewer round 2 (supersedes t_07753895): gate local-llm-hermes-desktop-rx7600 rev 2 — sha256 23467512…, 59,288 B
- t_07753895 ✓ done
Reviewer round 2: gate local-llm-hermes-desktop-rx7600 rev 2 (sha256 e4802dfc…, 59,117 B)
- t_0a951f57 ✓ done
Writer: apply reviewer FIX list to local-llm-hermes-desktop-rx7600 draft (3 should-fix + nits), then re-declare hash
subtasks 1/1
- ✓ done Reviewer round 2 (supersedes t_07753895): gate local-llm-hermes-desktop-rx7600 rev 2 — sha256 23467512…, 59,288 B
- t_c2bd20b9 ✓ done
Reviewer: gate the local-llm-hermes-desktop-rx7600 draft (sha256 6ee9f4a6…, 57,444 B)
subtasks 1/1
- ✓ done Writer: apply reviewer FIX list to local-llm-hermes-desktop-rx7600 draft (3 should-fix + nits), then re-declare hash
- t_e9e1e4db ✓ done
Research Hermes Desktop/Agent runtime requirements and candidate local LLMs
- t_4dc266fc ✓ done
Re-read prior local-LLM research PDF and extract hardware/runtime baseline
Recent actions 20
newest first · 20 of 50 events in the snapshot
- default heartbeat heartbeat
- default heartbeat heartbeat
- default worker_result worker finished
- default completed completed · Installed an unprompted workflow-closeout Discord notification rule in the two responsible profiles: `default` (workflow owner) and `discord-bot` (designated se
- default heartbeat heartbeat
- default commented commented
- default worker_bound worker_bound
- default owner_admitted owner_admitted
- default claimed claimed by worker
- default spawned worker spawned
- default worker_result worker finished
- default promoted promoted to ready
- default attached attached
- default completed completed · Audited the lab-update skills read-only and produced a 9-item change list (C1-C9) with exact file paths, edits and acceptance criteria in /opt/data/outputs-stag
- default heartbeat heartbeat
- default heartbeat heartbeat
- default worker_result worker finished
- default completed completed · Wrote the lab-workflow closeout note at /opt/data/notes/lab-update-workflow.md (5-step ship loop, 7 dated decisions, skills in play, commands, links, open quest
- discord-bot commented commented
- default heartbeat heartbeat
Sourced from agent_os in /lab/memory/graph.json — 24 tasks, 6 agents, last 20 of 50 events; snapshot . Rebuild with npm run graph.
Coverage: 117 nodes · 190 links — 13 pages · 53 files · 21 dirs · 6 agents · 24 tasks · 65 private entries hidden by policy; snapshot . The graph is generated at build time and does not poll.
How to read this
- Drag a node to rearrange it (it stays pinned until you press reset); drag empty space to pan.
- Click a node to highlight its links and neighbours and open its detail panel; Esc or an empty click clears it.
- Filters narrow by node type, search term, orphan status, or a local graph around the selection (depth 1–3).
- Agent OS is a snapshot, not a live feed: click an agent chip to filter its tasks and the activity log.