Switch the Live view's structure nodes + agent pawns from flat CircleGeometry
to lit SphereGeometry (MeshStandardMaterial with emissive + ambient/directional
lights), so they read as 3D balls from any orbit angle instead of going edge-on.
The additive glow sprites + particle effects are unlit, so the neon look is kept.
Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
Replaces the random per-id ellipsoid scatter with a fixed, deterministic brain
scaffold (frontend/src/components/world/brain.ts): two folded cerebral hemispheres
split by a longitudinal fissure (multi-octave noise = gyri/sulci), a cerebellum,
and a brain-stem — ~430 nodes wired into a k-nearest-neighbour edge network.
WorldCanvas Hierarchy renderer: the scaffold is one Points cloud + one LineSegments
(the dormant brain), with an activation field over the nodes. Agents seat into a
deterministic fixed region (agentRegion hash); their seat brightens on work and
fires activity-coloured signals along the real edges, which chain through the mesh
and light its pathways. Idle → a dim, gently-breathing brain. Same brain every
reload; orbit/pinch to see the 3D anatomy.
Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
Each agent carries a fireColor for its most recent activity — reasoning (purple),
tool (cyan), file op (coral), run (green) — derived in the engine from the touch
weight/node + reasoning events (now wired into the engine). The brain fires
signals in that colour, and the HUD shows a small colour key in Hierarchy, so you
can read WHAT KIND of work is lighting up the mesh.
Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
Clicking a neuron/pawn now focuses the Observe panel on that agent: a header with
its role + live status + a "← System" back button, then that agent's filtered
REASONING STREAM and TOOL CALLS. Feed items are tagged by agentId; the system
view shows the agent id per line. Wired via worldSel in Dashboard.
Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
A small bottom-left overlay on all three views: the view title + a one-line key
(Live = Gource tree, Flat = 2D graph, Hierarchy = brain mesh) plus live counts —
agents, currently-working, tokens/min and pending doors (from the telemetry feed).
Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
- brain edges now use per-vertex colors driven by endpoint activity, so a firing
neuron lights its pathways (gradient from the brighter end) while dormant
connections stay faint — the wiring visibly comes alive.
- neuron size now grows with cumulative activity (experience), so the agents that
do the most work read as the largest, most-developed neurons.
Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
Hierarchy (neural-brain) v2:
- structure nodes (org/company/team) join as dim support-neurons, so the brain
has body even with one or two agents (a real mesh, not a few dots).
- a resting-state shimmer: every neuron occasionally self-activates a touch, so a
dormant brain still breathes.
- agents brighten on working status AND on a recent touch (real activity), then
fire signals along pathways that chain through the support mesh.
Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
- Explosive sparks on file/project I/O: /api/world/live tags file-ish tools
(read/write/drive/vault/obsidian/edit/save) with touch weight 1; the engine
carries it as a node "burst" and the renderer fires a big fast particle blast.
- The formation tabs are now genuinely different views:
- Flat — the previous 2D React-Flow forest (WorldFlow), overlaid.
- Live — the WebGL Gource clone (glow sprites + sparks + trails + beams).
- Hierarchy — a NEW neural-brain view: agents are neurons in a dormant
two-hemisphere mesh; active agents fire signal particles along pathway edges
to their nearest neighbours (chaining), so the brain lights up as agents work.
- Render groups (gource/brain) + 3D particle system so signals travel in depth;
shared particle update; OrbitControls (rotate + pinch-zoom) across all WebGL views.
Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
Effects (matching Gource's technique — additive radial-gradient glow + particles,
not just post-processing):
- additive glow sprites on every node + pawn (a precomputed radial-gradient
texture, AdditiveBlending), scaled/brightened by heat — the soft Gource glow.
- a particle system (THREE.Points, additive, ring buffer): spark bursts on each
beam, ambient sparkle from hot nodes, and fading motion trails behind pawns.
- additive beams (brightness by life); UnrealBloom retuned (low threshold) so the
additive glow blooms.
Camera: PerspectiveCamera + OrbitControls — drag/one-finger to rotate (orbit),
pinch / wheel to zoom, two-finger / right-drag to pan; auto-frames until you grab
it. Tap-vs-drag raycast selection.
Formations now each behave distinctly:
- hierarchy — strong parent spring, low repulsion → clean structured tree (calm
pawns/effects).
- flat — spring to center + high repulsion → a spread peer mesh.
- live — balanced activity view with full pawns/beams/sparks/trails.
Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
frontend/src/components/observe/ObservePanel.tsx — a live mission-control view
driven entirely by the taxonomy feed (telemetry strip, doors awaiting approval,
agent statuses, and a rolling activity stream of real reasoning + tool calls;
the feed is ref-buffered + flushed on an interval so reasoning bursts don't
re-render per token). Rendered as the World slide-out's system view (replaces
the empty-state placeholder). With the run_events normalization, this shows
REAL agent reasoning, tool calls, and doors.
Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
/api/world/live now tails each active run's journal (run_events) and normalizes
the runner's existing events into taxonomy events — no runner change needed:
- text_delta -> agent.reasoning.delta (the live reasoning stream)
- step_started -> agent.tool.call + node.activity + world.touch on the tool
node (agents visibly converge on the tool they're using)
- approval_required -> door.request
Per-run seq cursor streams forward only (skips backlog on first sight). This
lights up REAL data for both the World view and the upcoming Observe surface.
Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
- /api/world/replay?hours= — reconstructs a sorted, timestamped taxonomy timeline
from the workspace's run history (agent_runs ⋈ sessions) for client playback.
- engine.clearWorldNodes() — drop transient world nodes/targets for a loop restart.
- WorldCanvas: a WorldClock control (Live ⇄ Replay) feeding the same engine —
play/pause, 1x/2x/4x speed, seek bar, loop. Live subscriptions detach in replay.
Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
- engine: when there are no service/event touch-nodes (a quiet feed), pawns roam
the real org/company/team structure so the view is always alive with zero
synthetic data; real touch-nodes take priority when present.
- /api/world/live: emit world.touch + node.activity per currently-running run
(agent_runs joined to sessions, state='running') so each working agent visibly
converges on its active-run node; fold running agents into working status;
telemetry.loops = active run count.
Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
Replaces the static React-Flow world forest with a real-time WebGL view of the
swarm working, plus the live event contract that feeds it.
Phase A — the live contract + feed:
- frontend/src/lib/live/taxonomy.ts — the 13-event Clawmates Event Taxonomy as
typed TS (the shared World/Observe contract).
- frontend/src/lib/live/useClawmatesLive.ts — native shared-singleton live client
(EventSource to /api/world/live, typed fan-out, replay-of-last-state to late
subscribers, refcounted, synthetic fallback so views are always alive).
- crates/cm-api/src/routes/world.rs — GET /api/world/live, authed + workspace-
scoped SSE emitting the taxonomy (real agent.status from live-container state,
a topology.update of the workspace's agents, telemetry with real doorsPending),
mirroring run_events_sse. normalize() seam documented for run_events->world.touch.
Phase B — the WebGL engine:
- frontend/src/components/world/engine.ts — Gource-inspired force-directed model:
org/company/team tree (sibling repulsion + parent spring + friction), agent
pawns that converge on the touched node and beam it, world nodes that glow with
heat and fade when idle. Framework-agnostic (renderer-independent) state+math.
- frontend/src/components/world/WorldCanvas.tsx — three.js scene (ortho cam,
UnrealBloom), render loop syncing engine state, camera auto-fit, HTML labels,
raycast click->select, Hierarchy/Flat/Live formation switch.
- Dashboard.tsx: swap <WorldFlow/> -> <WorldCanvas/> at the world-tier seam
(shared claw-tier pieces untouched; WorldFlow.tsx kept for now).
- Adds three (+ @types/three).
Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
Decouples "many users" + "many server replicas" from "many machines" so the
platform is tenant-isolated and horizontally safe on the current single node.
- Per-signup workspaces (cm-auth): a new hosted-identity sign-in provisions and
owns its own workspace instead of joining the first. Config-gated by
auth.per_signup_workspace (default off); concurrent first-logins serialized by
a per-subject advisory lock so no duplicate workspaces.
- Terminal tickets in Postgres (migration 0016, hashed, single-use): any replica
can redeem a ticket minted by another. Drops the in-process ticket map.
- Container registry in Postgres (migration 0017, agent_containers): Terminal
and Sandbox managers resolve an agent's container through a shared registry,
so a 2nd replica reuses it instead of spawning a duplicate. node_id recorded
as 'local' (Phase 2 hook). Boot reconcile removes only true orphans, so
terminals now survive a redeploy (tmux sessions resume).
- Per-workspace quotas (cm-api/quota.rs): plan-tier caps on agents + live
containers, enforced at agent create + terminal spin-up (reconnects allowed),
returned as HTTP 402. New GET /api/quota surfaces usage vs limits.
Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
Terminal app (xterm ⇄ WebSocket ⇄ per-agent themed container):
- zsh + oh-my-zsh + powerlevel10k image (agent-terminal), runs as uid 65532 to
share read-write ownership of the file-drive volume with the server.
- Interactive PTY in cm-sandbox (bollard exec tty/attach + resize) + a
TerminalManager; ticket-authed WS bridge routed straight to the backend via a
Traefik PathRegexp(/ws) rule. MOTD greets the user by name.
- tmux resumable sessions; multi-tab (one tmux session per tab, same container),
drag-to-reorder, rename, and a Save that persists named tabs to the server
(terminal_tabs, migration 0014) so they survive logout / a new device.
- Files drives mounted per-agent (subpath) at ~/drives/{documents,received,
shared}; a reconciler keeps the Files app's index in sync with terminal writes.
Storage moved to a shared `filedata` volume (CLAWMATES_STORAGE__DATA_DIR).
Obsidian vault (a markdown "second brain" per agent):
- New `vault` FileDrive (migration 0015) mounted into the terminal at ~/obsidian;
a file-content read route; a purple Obsidian tile + a vault viewer app.
Computer UI:
- Draggable computer-panel width (min = phone preset) keeping the size presets.
- Green Terminal glyph, "Claw Chat" → "Chat", colored gradient-outline app icons.
- Agent page: avatar↔activity-grid spacing + larger, uniform section fonts with
colored section-tinted tag chips.
Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
Replaces the approximation with an exact port of Clawmates Dashboard.html:
- breadcrumb Acme Org / Acme Corp / Growth Team / <claw> (active-tier highlight)
- ORG/CO/TEAM/CLAW tier rail; clicking a tier shows the example structure
(org→3 companies, company→4 teams, team→6 agents)
- morphing topology canvas with all 6 patterns (hub-spoke/pipeline/ring/mesh/
swarm/debate) + exact POS/LNK tables + animated links
- company grid canvas (team cards) + org radial canvas (company nodes)
- claw view: left compartment panel + ANATOMY canvas (Skills/Personality/Memory/
Tools·Doors/Capabilities/Safety) + the agent "Computer" slide-out
(apps / now-running / dock), matching the doc
- exact status colours, gradients, animations, status bar
Example workspace data baked in (as in the doc) so every tier is populated;
real-workspace wiring is a follow-up layer.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Login landed on an empty "Pick a claw" stub — the integrated dashboard from the
design comp was never assembled as the home. Now `/` IS the dashboard.
- public/sw.js: cache v2→v3; RegisterServiceWorker reloads once when a new
worker activates, so deploys are picked up without a manual hard-refresh
- (workspace)/layout: ShellChrome renders the dashboard bare on "/" (it's
self-contained) and the shared TopBar/LeftRail/StatusBar on every other route
- components/dashboard: Dashboard (state machine + data) — top bar w/ breadcrumb,
ORG/CO/TEAM/CLAW tier rail, tier-aware context list, TopologyCanvas (6-mode
view-as selector, generalized layouts), and the agent "computer" slide-out
(cm-fade) wrapping the existing ComputerPanel; status bar
- motion.css: cm-fade keyframe
Wired to existing endpoints (/api/teams|companies|orgs, /api/structure/*,
/api/team/claws, /api/structure/stats, + ComputerPanel's apps/runtime-config/
routines). Defaults to the most-recent team; claw click opens the slide-out.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
The token redesign couldn't touch static rasters/config — these were still the
old claw mark + old coral. Regenerate favicon (icon.svg + icon.png) and PWA
icons (192/512) as the node-mesh glyph; manifest theme_color #f96565→#ff6f61,
background #121212→#08080a; refresh the description/tagline.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
The "+" deploy flow now opens a scope selector (Single / Team / Company-soon /
Org-soon). Single = the existing CreateClawForm. Team = TeamWizard: pick a
baseline topology (from the catalog + role distribution), set size, auto-staff
editable claw cards (name / role / model / persona), preview the topology
graph, then POST /api/teams → land on the Team page.
Team page (/teams/[id]): topology SVG + claw roster (each links to its chat) +
a Run panel that drives the team on the durable runner with live SSE progress
(reuses the topology-run streaming). /teams list page + a Teams nav entry.
lint + typecheck + next build clean.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
A team = a baseline topology staffed with real claws. Migration 0010 (teams +
team_members node→claw bindings) + cm-db repo/teams.rs. runtime_provision.rs
turns a claw into a live runtime agent claw_<id> via the synced gateway config
API (#7468): create agent + bind model_provider (mapped from chosen model) +
risk_profile=toolfree + clawmates_door bundle — atomic, immediately drivable.
routes/teams.rs: POST /api/teams (create claws + provision + build(kind,roles)
+ bind node.attrs["agent"]=claw_<id> + persist), GET /api/teams[/{id}],
POST /api/teams/{id}/run (enqueue a durable run of the team graph — reuses the
topology worker + SSE). v1 persona = topology role via the prompt builder; the
claw's system_prompt stays its chat identity.
Spike confirmed: runtime agent provisioning works; IDENTITY.md persona works
for API models (Gemini/Groq), masked by CLI models (Claude/Kimi Code). 16
cm-api tests + provision unit tests pass, clippy clean.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
The §15 governor can now be judged by a ZeroClaw runtime agent instead of a
server-side registry/API model. ZeroClawDriveExecutor::judge drives an agent
with the governor prompt and parses ALLOW/DENY (fail-open). mcp_door routes to
it when CLAWMATES_JUDGE_MODEL=runtime:<alias>.
This unblocks Kimi-as-judge with NO Kimi Platform key: Kimi runs on the
membership via kimi_cli, so CLAWMATES_JUDGE_MODEL=runtime:judge_kimi makes the
coding agent the governor. (Same path works for any subscription-only model.)
cm-runtime re-exports judge_model. clippy clean.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Gemini's project-wide key-restriction 403 is cleared (key restricted to the
Generative Language API), so Gemini rejoins the grid. Full 2×4 ranking
(pipeline/debate × Claude/GLM-4.7/Gemini/heterogeneous), judge-ranked 1-8:
D2 > P2 > P3 > P1 > D3 > PG > DG > D1.
Adds a third finding: model strength is task-dependent — Gemini-2.5-flash
placed 2nd on the §7.3 concision brief but 6th-7th here on reasoning
("generic list", "idea repeated"): concise-but-shallow. Reinforces that the
optimal config is a joint choice over {topology × per-role model × task}.
Gemini cells ran via the durable async run path (enqueue → worker → poll).
Co-Authored-By: Claude Opus 4.8 <[email protected]>
The §15 door's email_send queues to `outbox` but nothing delivered it. Add a
real transport: cm-db outbox repo (list_queued/mark_sent/mark_failed) + a
cm-runtime drainer — an EmailSender trait (testable), a lettre STARTTLS
LettreSender, drain_once (queued -> sent/failed), and spawn_drainer wired into
the server beside the scheduler/sweeper/topology-worker.
Config-gated: inert (logs "outbox delivery DISABLED") until CLAWMATES_SMTP_*
is set, so it ships safely before credentials exist. The agent never holds the
SMTP credential — it only writes to outbox through the gated door; the server
owns the transport.
NOTE: live delivery is still credential-blocked — Migadu's API can't send
(SMTP-only) and the admin token is invalid; no SMTP creds exist. The transport
is built + tested (drain_marks_sent_and_failed via a mock sender); set
CLAWMATES_SMTP_* to go live with zero further code. clippy clean.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
#2 GLM judge (closes#114): registry NamedProvider gains a `format` field;
build_provider_registry builds an AnthropicProvider for format="anthropic".
GLM's coding/OpenAI endpoint is ToS-throttled for raw SDK, but its Anthropic
endpoint (api.z.ai/api/anthropic) accepts raw API calls (verified x-api-key
-> glm-4.7), so CLAWMATES_JUDGE_MODEL=glm:glm-4.7 routes the door governor /
topology judge through GLM with no runtime-routing. (Kimi-as-judge still needs
a Platform key — coding key is agent-only.)
#3 run-control: POST /api/topology-runs/{id}/cancel (workspace-scoped,
queued/running only); the worker honors it at the step boundary (checks
current_status in the checkpoint callback) and won't clobber a cancel with
`failed`. Frontend Run tab gains a Cancel button and clickable recent runs
that deep-link into a live/replayed stream (SSE replays from checkpoint).
cm-* tests (incl. new cancel test) + clippy + frontend lint/typecheck green.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Backend: GET /api/topology-runs/{id}/events streams Server-Sent Events by
tailing the durable per-step checkpoint the worker already writes — a `step`
event per newly-completed step (replayed on connect so reload/reconnect
re-attaches), then a terminal `done` event with the final output/error. Each
step carries its index as the SSE id, so the browser's Last-Event-ID resumes
without duplicates on reconnect. No new table, no worker change — reuses the
checkpoint; avoids run_events' agent_runs FK.
Frontend: the Run tab now opens an EventSource (through the same-origin proxy,
which adds the bearer) instead of polling — appending steps as they stream and
finalizing on `done`. One streaming connection, lower latency, auto-resume.
cm-api builds + clippy clean; frontend lint + typecheck + next build clean.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
New TopologyRun component: pick a topology + roles + task, enqueue via
POST /api/topologies/run (202), then poll GET /api/topology-runs/{id} for
live turn-by-turn progress — rendering the per-step checkpoint while running
and the final RunRecord (+ final output) when completed, with a status badge
and recent-runs list. Added as the middle "run" tab in the workbench.
Since the work runs server-side as a durable job, a long-horizon run survives
navigation/restarts; the UI just re-attaches by polling. lint + typecheck +
next build clean. (Also fixed a pre-existing set-state-in-effect lint error in
TopologyCompare surfaced by the Next 16 toolchain.)
Co-Authored-By: Claude Opus 4.8 <[email protected]>
POST /api/topologies/run now ENQUEUES a durable job and returns 202
{run_id, status:queued} instead of executing inside the HTTP request — the
prerequisite for long-horizon runs (no client/proxy/LB timeout, survives
restarts).
topology_worker: a spawned loop that requeues stale running jobs, claims the
next queued one (CAS via FOR UPDATE SKIP LOCKED), drives it through
execute_resumable, and checkpoints RunProgress after every step; on crash the
stale sweep requeues it and the next claim resumes from the last checkpoint.
Wired into server startup beside the scheduler + resume sweeper.
GET /api/topology-runs/{id} now reports lifecycle status/kind/error/checkpoint
+ the result blob (kept the `comparison` field name for back-compat with the
compare UI; null until completed). list_runs includes status + kind.
Tests: durable lifecycle (enqueue→claim→checkpoint→complete) + stale-requeue
resume, both green; p0 endpoints (compare path) unchanged. 13 + 2 tests pass,
clippy clean.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Split execute() into a thin wrapper over new execute_resumable(), which starts
from a prior RunProgress checkpoint (completed step count + outputs + records +
metrics) and invokes an async on_step callback after each newly completed step.
The step plan is re-derived from the graph (planners are deterministic), so only
completed outputs need persisting; the callback owns persistence, keeping the
orchestrator storage-agnostic. RunProgress + the journal types (StepRecord,
RunMetrics, StepPhase, GatedAction) gain Deserialize for JSONB round-trip.
New test: resume-from-checkpoint runs only the remaining steps and reproduces
the full run's output. 14 tests pass, clippy clean.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Evolve topology_runs into a durable-job table (migration 0009): status state
machine (queued/running/completed/failed/cancelled), kind, input graph,
per-step checkpoint, error, last_event_id, timestamps. comparison becomes
nullable (the result blob, absent until completion). Back-compat: existing
rows default to completed/compare.
cm-db repo gains the durable-job ops: enqueue_run, claim_next_queued (CAS via
FOR UPDATE SKIP LOCKED), checkpoint, complete, fail, requeue_stale (resume
sweep), and status(). Regenerated .sqlx cache. Also fix two pre-existing test
RuntimeConfig literals missing the providers field (from the registry work).
Co-Authored-By: Claude Opus 4.8 <[email protected]>
The GEMINI_API_KEY was restricted (fixing the prior 403 enforcement), so
Gemini-2.5-flash rejoins the per-role leaderboard. Full ranking:
GLM-4.7 > Gemini-2.5-flash > Claude > Kimi > GLM-5.2. Flagship GLM-5.2 still
last on the concision-weighted brief.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Crosses pipeline x debate with 3 model casts (Claude, GLM-4.7, heterogeneous)
on a fixed decision task, judge-ranked 1-6. Central finding: the topology x
model interaction is real and non-monotone — debate amplifies the strongest
model (GLM-4.7: 2nd->1st) but degrades others (Claude: 3rd->last). The optimal
configuration is a joint choice over {topology x per-role model}, which neither
a topology-only nor model-only study can surface. 21 turns, sequential, within
subscription quota caps. Safety invariant holds across all cells.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Single-role drafter on a fixed concision-weighted brief, one turn per backend,
judge-ranked: GLM-4.7 > Claude > Kimi > GLM-5.2. Notable inversion — the
flagship GLM-5.2 ranked last by over-writing the 25-word brief, underscoring
that the best model is role/task/rubric-dependent. Adds the worker_glm5
(glm-5.2 Opus-class) tier to the example config. Gemini excluded this run
(Google June-19 API-key-restriction 403; infra, not model).
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Adds the Phase C result: five backends (Claude/Kimi/GLM/Gemini/Groq) as
interchangeable role-agents, per-node model selection via node.attrs["agent"],
and a 3-config drafter->refiner sweep judged for quality. Heterogeneous
GLM->Gemini won over the all-Claude baseline — an existence proof of
"which model wins which role in which topology." Notes quota-bounded sweep
scope as future work.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
TurnRequest gains an optional `agent` sourced from the graph node's
attrs["agent"]. The ZeroClaw executor binds a node to that alias directly
when present, falling back to the role→alias map otherwise. This lets a
single POST /api/topologies/run specify a different model per role
(heterogeneous topologies) entirely in the graph JSON — no server
ZEROCLAW_AGENT_MAP change or recreate per configuration, which makes
quota-frugal model×role sweeps practical.
cm-orchestrator 13 + cm-api topology_exec 4 tests pass, clippy clean.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Gemini via the built-in `gemini` API family (gemini-2.5-flash default,
gemini-3-pro-preview flagship) and Groq via `groq` — both wired as topology
worker aliases (worker_gemini, worker_groq), no image rebuild. Documents the
Phase C recipe: map semantic roles to per-model aliases via the server
ZEROCLAW_AGENT_MAP, then POST /api/topologies/run.
Five subscription/API backends now selectable per role: Claude, Kimi, GLM,
Gemini, Groq. Proven live on gw-04 — a 3-stage pipeline ran researcher=GLM-4.7
-> analyst=Kimi-K2 -> writer=Gemini-2.5-flash end-to-end (RunRecord persisted).
Quota-care notes added (5h/weekly caps, low concurrency -> prefer sequential
pipelines, short prompts).
Co-Authored-By: Claude Opus 4.8 <[email protected]>
GLM Coding Plan has no CLI of its own and ToS-throttles raw SDK/API access
("1113 Insufficient Balance"); it is served only inside officially supported
tools, and Claude Code is the primary one. So GLM rides the existing
claude_cli provider with a z.ai env override (ANTHROPIC_BASE_URL=
api.z.ai/api/anthropic + ANTHROPIC_AUTH_TOKEN) — no new provider, no image
rebuild. Fresh HOME + CLAUDE_CODE_OAUTH_TOKEN="" so the container's cached
Claude subscription doesn't override the routing; key via $NAME indirection.
Deployed & verified on gw-04: worker_glm answers "I am GLM-4.7, built by
Zhipu AI" over /ws/chat. Three subscription models now co-reside on one
runtime — Claude (coordinator), Kimi (worker_kimi), GLM (worker_glm).
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Runtime image now installs Node 22 (nodesource) + both agent CLIs —
@anthropic-ai/claude-code and @moonshot-ai/kimi-code — so the zeroclaw
daemon can spawn `kimi -p` for the new kimi_cli provider on a Kimi
membership, parallel to `claude -p` (claude_cli). Kimi needs Node >=22.19;
the image was on Node 18.
Documents the kimi_cli worker recipe in agent.config.example.toml (the
coding endpoint is User-Agent gated → drive Kimi through its own CLI, not a
raw OpenAI-compat client) and keeps GLM via the registry family.
Deployed & verified on gw-04: coordinator (claude_cli) answers "Anthropic
built me", worker_kimi (kimi_cli) answers "I am Kimi, built by Moonshot AI"
— heterogeneous Claude + Kimi on one runtime, over /ws/chat.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Phase A wiring (config-only): server registry kimi provider now targets the
Kimi Code coding endpoint (api.kimi.com/coding/v1, model kimi-for-coding,
subscription-billed via a Kimi Code Console key). Runtime example documents a
custom.kimi worker provider (OpenAI-compatible coding endpoint) + GLM, selectable
per topology role via ZEROCLAW_AGENT_MAP. Activates on KIMI_API_KEY; no rebuild.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
ZeroClaw already supports glm + moonshot (Kimi) as model providers; document the
pattern to run a topology worker on them (env model+key, agent on the alias,
role->alias via ZEROCLAW_AGENT_MAP).
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Adds a multi-provider registry so judges and topology execution can run on
providers beyond the default. cm-config gains [[llm.providers]] (name, base_url,
api_key_env) — all OpenAI-compatible (GLM, Kimi/Moonshot). The server builds an
Arc<dyn LlmProvider> per entry (OpenAiCompatProvider) keyed by name; a missing
key is skipped with a warning, not a boot failure. cm-runtime RuntimeConfig
carries a ProviderRegistry; Runtime::resolve_provider("<name>:<model>") selects a
registry provider (else the default). The door governor (Runtime::judge) and the
topology compare endpoint both resolve through it, so CLAWMATES_JUDGE_MODEL and
CLAWMATES_TOPOLOGY_EXEC_MODEL accept "glm:glm-4.6" / "kimi:kimi-k2".
Co-Authored-By: Claude Opus 4.8 <[email protected]>
