User pushback: "транслирует в онлайне какие агенты создаются? основное
окно агента, а дальше при запусках появляются новые ветки, мы показываем
в онлайне как агенты собираются и работают"
Earlier `kei-graph-export` rendered the static SUBSTRATE (all 581 atoms,
catalog-style). User wanted the LIFECYCLE: orchestrator at center, every
new agent as a fading-in branch, every tool call as a pulse, every
completion as a fade-out. TTL = until done; pure online, no history
accumulation per user direction.
Three-layer architecture, all conforming to schema /tmp/agent-events-schema.md:
LAYER 1 — Event emitters (4 hooks)
hooks/agent-event-spawn.sh PreToolUse:Agent → agent_spawn event
hooks/agent-event-done.sh PostToolUse:Agent → agent_done event
(parses STATUS-TRUTH MARKER for outcome,
computes cost_usd from token×pricing table)
hooks/tool-use-event.sh PreToolUse:Bash|Read|Edit|Write|Grep|Glob|NotebookEdit
→ tool_use event
hooks/skill-record.sh EXTENDED — second emit step writes skill_use
event in addition to existing kei-ledger
record-skill call
All 4 are POSIX /bin/sh, defensive (never block, exit 0), bypass via
KEI_EVENTS_BYPASS=1. Append-only JSONL to
~/.claude/memory/agent-events.jsonl.
Smoke: 4 synthetic invocations cover spawn/done/tool/filter cases.
LAYER 2 — kei-graph-stream Rust daemon
_primitives/_rust/kei-graph-stream/ (~480 LOC, 5 files + 1 test)
- Tails events.jsonl every 200ms (poll-based, no notify dep).
- Parses each event, updates AliveState (insert on spawn, remove on done).
- Broadcasts {"type":"event","data":<event>} to all WebSocket clients.
- On client connect: sends {"type":"snapshot","alive":[...]} first.
- Heartbeat: {"type":"ping"} every 30s.
- axum 0.7 + ws feature (already in Cargo.lock via kei-cortex).
- Bypass: KEI_GRAPH_STREAM_BYPASS=1.
Bound to 127.0.0.1:8201 (loopback only). Endpoints:
GET /stream → WebSocket upgrade
GET /health → "kei-graph-stream alive"
4 unit + 1 integration test. cargo build clean.
Installed binary: ~/.cargo/bin/kei-graph-stream
Launchd plist: io.keisei.graph-stream (RunAtLoad, KeepAlive)
Loaded as PID 52678, /health 200 OK verified.
LAYER 3 — live-graph.html (single-file frontend)
~/Projects/lbm-graph-viz/live-graph.html (~464 LOC, self-contained)
- SVG full-viewport, dark #0f172a, CSS grid background.
- Pinned center node "main" (orchestrator), gold #fbbf24, glowing.
- Agents radiate via D3 force-simulation; color-by-model
(sonnet=green, opus=red, haiku=blue, default=gray).
- On agent_spawn: fade-in 300ms, edge from main to new node.
- On tool_use: pulse on agent node (r 8→12→8 over 400ms) +
floating tool name label fades 800ms.
- On agent_done: outcome-color flash → fade-out 800ms → remove.
- WebSocket client: ws://127.0.0.1:8201/stream, exponential-backoff
reconnect (1s→30s).
- Top-right status badge: ● connected | ○ reconnecting | ✕ disconnected.
- Bottom counters: alive / spawned / tool calls / done / last event age.
- No build step. D3 v7 from CDN. Pure HTML+JS+CSS.
End-to-end smoke (this machine, just now):
- daemon health 200 OK
- hook injected agent_spawn → daemon broadcasts → AliveState=1
- hook injected agent_done → daemon broadcasts → AliveState=0
- frontend file syntax-checked clean
What this does NOT do (deferred, by user direction "это онлайн"):
- History persistence — agents who finished are GONE from the graph.
Per-session log remains in events.jsonl + sleep-sync if user wants
to consult later, but the live view is RIGHT NOW only.
- Sub-agent attribution beyond "main" — orchestrator-direct tool calls
show on the orchestrator node. Sub-agent's internal tool calls would
need session-id correlation; current schema has agent_id="main"
placeholder for non-Agent tool calls.
- Replay mode — no time-scrubber. Possible follow-up if useful.
- Auth on WebSocket — bound to 127.0.0.1 only. Local-only by design.
=== STATUS-TRUTH MARKER ===
shipped: functional
stubs: 0
cargo-check: PASS
behaviour-verified: yes
follow-up-required:
- Sub-agent tool-call attribution (correlate session_id chain)
- Replay mode with time scrubber (if user finds use)
- Tool aggregator nodes ("Bash bucket" with N) instead of per-agent pulses
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
232 lines
9.4 KiB
TOML
232 lines
9.4 KiB
TOML
[workspace]
|
|
resolver = "2"
|
|
members = [
|
|
"kei-ledger",
|
|
"kei-migrate",
|
|
"kei-changelog",
|
|
"ssh-check",
|
|
"firewall-diff",
|
|
"mock-render",
|
|
"visual-diff",
|
|
"tokens-sync",
|
|
"kei-memory",
|
|
"kei-conflict-scan",
|
|
"kei-refactor-engine",
|
|
"kei-graph-check",
|
|
"kei-store",
|
|
# v0.14 LBM port — 10 new MCP-core primitives
|
|
"kei-router",
|
|
"kei-sage",
|
|
"kei-task",
|
|
"kei-chat-store",
|
|
"kei-crossdomain",
|
|
"kei-search-core",
|
|
"kei-content-store",
|
|
"kei-social-store",
|
|
"kei-curator",
|
|
"kei-auth",
|
|
# v0.15 artifact handoff pipeline
|
|
"kei-artifact",
|
|
# v0.18 exobrain CLI
|
|
"keisei",
|
|
# v1 substrate — local web wizard for scaffolding atoms (Stream A)
|
|
"kei-forge",
|
|
# v1 substrate — atom invocation runtime + schema linter (Stream D)
|
|
"kei-runtime",
|
|
# Hosted Sleep substrate — 12 traits + DNA + plugin registry (impls in sibling crates)
|
|
"kei-runtime-core",
|
|
# Hosted Sleep Wave 2 — bare-metal SSH ComputeProvider impl
|
|
"kei-compute-baremetal",
|
|
# v1 substrate — shared atom discovery + frontmatter + safe path (Stream E)
|
|
"kei-atom-discovery",
|
|
# agent substrate v1 — phase 3 runtime (Capability trait + registry + compose/spawn/verify)
|
|
"kei-agent-runtime",
|
|
# agent substrate v1 — phase 3 hook-protocol CLI adapter
|
|
"kei-capability",
|
|
# v0.24 unification — unified VPS provisioner (supersedes provision-{hetzner,vultr}.sh)
|
|
"kei-provision",
|
|
# Convergence Layer A — schema-driven verb-template engine for SQLite-CRUD stores
|
|
"kei-entity-store",
|
|
# v1 substrate — atom DAG pipe runtime (topo-sort + JSON piping between atoms)
|
|
"kei-pipe",
|
|
# v1 substrate — deterministic result cache for pure (query/transform) atoms
|
|
"kei-cache",
|
|
# agent substrate v1 — automation envelope: prepare + ledger fork + verify
|
|
"kei-spawn",
|
|
# agent substrate v1 — reconstruct spawn from DNA (ledger row + task.toml + recompose)
|
|
"kei-replay",
|
|
# v0.29 Wave 13 — structural JSON diff primitive (RFC 6902 subset add/remove/replace)
|
|
"kei-diff",
|
|
# v0.29 Wave 13 — durable task scheduler (cron / at / interval) metadata primitive
|
|
"kei-scheduler",
|
|
# v0.29 Wave 13 — filesystem watcher primitive (thin notify wrapper, sync API)
|
|
"kei-watch",
|
|
# v0.30 Wave 14 — bio-inspired pruning (retire idle agents)
|
|
"kei-prune",
|
|
# v0.30 Wave 14 — federated marketplace discovery stub
|
|
"kei-discover",
|
|
# v0.30 Wave 14 — taxonomy graph + lineage stdout visualizer
|
|
"kei-brain-view",
|
|
# v0.30 Wave 14 — whole-brain tar.zst export/import
|
|
"kei-hibernate",
|
|
# v0.30 Wave 14 — ed25519 creator attestation
|
|
"kei-ledger-sign",
|
|
# v0.31 Wave 15 — managed git worktree + ledger lifecycle (fork/collect/gc/rescue)
|
|
"kei-fork",
|
|
# v0.34 Wave 17 — SSoT for DNA format + shared substrate types
|
|
"kei-shared",
|
|
# v0.32 Wave 15 — read-only DNA adjacency/cluster/precedent over kei-ledger
|
|
"kei-dna-index",
|
|
# Pet UI v1 — persona manifest parse/validate + Ed25519 identity + overlay renderer
|
|
"kei-pet",
|
|
# v0.37 Wave 20 — local HTTP daemon (axum) exposing cortex state for web UI at keisei.app
|
|
# v0.39 Wave 25 — longitudinal user-frustration matrix (regex + byte n-gram firmware classifier)
|
|
"frustration-matrix",
|
|
# Wave 54 — per-user online learning loop on top of frustration-matrix (sleep Phase 0 + install bootstrap)
|
|
"kei-frustration-loop",
|
|
# Wave 26.5 — universal external skill-format importer (OpenClaw/Cline/Cursor/Claude/Kimi)
|
|
"kei-skill-importer",
|
|
# Wave 45 — dev-hub bundle (project dashboard data sources)
|
|
"kei-projects-index",
|
|
"kei-projects-watcher",
|
|
# Wave 46 — Google Drive → Forgejo one-shot importer (folder classifier)
|
|
"kei-gdrive-import",
|
|
# Wave 48 — content protection SSoT (regex matrix scanner / substituter)
|
|
"kei-leak-matrix",
|
|
# Wave 51 — research output → action pipeline (parses MASTER-REPORT.md, emits kei-spawn task.toml)
|
|
"kei-decision",
|
|
# Wave 52 — UNIVERSAL decomposition layer: 5 MD-format parsers (research/audit/sleep/architecture/new-project)
|
|
"kei-decompose",
|
|
# A1.1 — foreign project ingestion runtime (clone → walk → identify modules → stage for downstream atomars)
|
|
"kei-import-project",
|
|
# Wave 53 — UNIVERSAL block identity registry (primitive/skill/rule/hook/atom DNA in SQLite)
|
|
"kei-registry",
|
|
# Wave 55 — UNIVERSAL model registry + selector (verified pricing 2026-04-28)
|
|
"kei-model",
|
|
# P1.2 daytona
|
|
"kei-backend-daytona",
|
|
# P0.2 trajectory export
|
|
"kei-export-trajectories",
|
|
# Phase 3.1 — Hermes/agentskills.io SKILL.md format primitives
|
|
"kei-skills",
|
|
# P4.1 gateway — multi-platform messaging unified (Telegram/Discord/Slack/CLI/WhatsApp/Signal/Generic)
|
|
"kei-gateway",
|
|
# P4.2 scheduler — Hermes-equivalent cron/at/interval with JSON persistence
|
|
# NOTE: distinct from existing `kei-scheduler` (SQLite metadata DB). See
|
|
# `kei-cron-scheduler/Cargo.toml` for the rationale on coexistence.
|
|
"kei-cron-scheduler",
|
|
# Wave 56 — Mac hardware/OS/tooling detector (foundation for local LLM stack: ollama/llamacpp/mlx)
|
|
"kei-machine-probe",
|
|
# Wave 57 — Ollama HTTP adapter (localhost:11434, streaming, 5 subcommands)
|
|
"kei-llm-ollama",
|
|
# Wave 58 — llama.cpp shell-out adapter (no FFI, llama-cli/llama-server, 5 subcommands)
|
|
"kei-llm-llamacpp",
|
|
# Wave 59 — Apple MLX adapter (mlx_lm shell-out, macOS Apple Silicon only, hard platform gate)
|
|
"kei-llm-mlx",
|
|
# Hosted Sleep Wave 4 — MLX LlmBackend bridge (Apple Silicon only)
|
|
"kei-llm-bridge-mlx",
|
|
# Hosted Sleep Wave 2 — Vultr ComputeProvider impl
|
|
"kei-compute-vultr",
|
|
# Hosted Sleep Wave 2 — Linode ComputeProvider impl
|
|
"kei-compute-linode",
|
|
# Wave 60 — UNIVERSAL local-LLM backend selector (glues W55+W56+W57+W58+W59 → picks Mlx/LlamaCpp/Ollama)
|
|
"kei-llm-router",
|
|
# Hosted Sleep Wave 1 — systemd ServiceManager impl
|
|
"kei-svc-systemd",
|
|
# RULE 0.17 — atomar time-logger (Wave 10)
|
|
"kei-tlog",
|
|
# Per-LLM-call token + cost observability store (cortex token tracker, Phase D sleep-report)
|
|
"kei-token-tracker",
|
|
# Cross-window agent heartbeat (auto: redis | sqlite)
|
|
"kei-ping",
|
|
# Hosted Sleep Wave 2 — DigitalOcean ComputeProvider impl (Wave 5 redo)
|
|
"kei-compute-digitalocean",
|
|
# Hosted Sleep Wave 5 — git providers (Gitea/Forgejo/GitLab/Bitbucket)
|
|
"kei-git-gitea",
|
|
"kei-git-forgejo",
|
|
"kei-git-gitlab",
|
|
"kei-git-bitbucket",
|
|
# Hosted Sleep Wave 6 — memory backends (sled/redis/postgres/sqlite)
|
|
"kei-memory-sled",
|
|
"kei-memory-redis",
|
|
"kei-memory-postgres",
|
|
"kei-memory-sqlite",
|
|
# Wave 7 — Google OAuth 2.0 + OIDC AuthProvider
|
|
"kei-auth-google",
|
|
# Wave 7 — Apple Sign-In AuthProvider
|
|
"kei-auth-apple",
|
|
# Wave 7 — Magic-Link email AuthProvider (HMAC-signed token, stateless)
|
|
"kei-auth-magiclink",
|
|
# Wave 7 — WebAuthn passkey AuthProvider
|
|
"kei-auth-webauthn",
|
|
# Wave 8 — Telegram Bot API NotifyChannel (sendMessage with HTML parse_mode + severity emoji)
|
|
"kei-notify-telegram",
|
|
# Wave 8 — Discord webhook NotifyChannel
|
|
"kei-notify-discord",
|
|
# Wave 8 — Slack incoming-webhook NotifyChannel
|
|
"kei-notify-slack",
|
|
# Wave 8 — SMS via Twilio Programmable Messaging
|
|
"kei-notify-sms",
|
|
# Wave 9 — VPN/network backends
|
|
"kei-net-wireguard",
|
|
"kei-net-openvpn",
|
|
"kei-net-ipsec",
|
|
# Cortex stack — local AI runtime (daemon + TUI + MCP server)
|
|
"kei-cortex",
|
|
"kei-tty",
|
|
"kei-mcp",
|
|
# SQL ↔ TypeScript schema drift detector
|
|
"kei-db-contract",
|
|
# Live runtime-graph exporter (registry + ledger → D3 space fragment)
|
|
"kei-graph-export",
|
|
# Live agent-events.jsonl tail → WebSocket stream (kei-graph-stream daemon)
|
|
"kei-graph-stream",
|
|
]
|
|
|
|
[workspace.package]
|
|
edition = "2021"
|
|
rust-version = "1.77"
|
|
|
|
[workspace.dependencies]
|
|
clap = { version = "4", features = ["derive"] }
|
|
serde = { version = "1", features = ["derive"] }
|
|
serde_json = "1"
|
|
serde_yaml = "0.9"
|
|
serde_yaml_ng = "0.10"
|
|
sha2 = "0.10"
|
|
image = { version = "0.25", default-features = false, features = ["png"] }
|
|
regex = "1.10"
|
|
rusqlite = { version = "0.31", features = ["bundled"] }
|
|
chrono = { version = "0.4", default-features = false, features = ["clock"] }
|
|
anyhow = "1"
|
|
thiserror = "1"
|
|
tempfile = "3"
|
|
toml = "0.8"
|
|
flate2 = "1"
|
|
walkdir = "2"
|
|
pretty_assertions = "1"
|
|
# Shared async + HTTP deps (Waves 31/32/33/35/36 etc — kei-tty / kei-router / etc)
|
|
tokio = { version = "1", features = ["rt-multi-thread", "macros", "signal", "net", "time", "process", "fs", "io-util", "sync"] }
|
|
tokio-stream = "0.1"
|
|
futures = "0.3"
|
|
reqwest = { version = "0.12", features = ["json", "stream", "multipart", "rustls-tls"], default-features = false }
|
|
bytes = "1"
|
|
async-trait = "0.1"
|
|
wiremock = "0.6"
|
|
# Wave 38a — daemon integration deps (PTY for /term WS endpoint, tungstenite for axum WS)
|
|
portable-pty = "0.8"
|
|
tokio-tungstenite = "0.29"
|
|
# Wave 44a — tool-sandbox hardening (shell tokenization, URL parsing, bounded LRU)
|
|
shell-words = "1"
|
|
url = "2"
|
|
lru = "0.12"
|
|
# Wave 44b — atomic O_NOFOLLOW openat for symlink-safe writes (tool_apply.rs)
|
|
nix = { version = "0.29", default-features = false, features = ["fs"] }
|
|
# A2.1 — kei-import-project trait pattern matcher (syn AST parsing)
|
|
syn = { version = "2", features = ["full"] }
|
|
|
|
[profile.release]
|
|
opt-level = "z"
|
|
lto = true
|
|
strip = true
|
|
codegen-units = 1
|