An honest separation of what Kazma does today from what is planned, aspirational, or partially wired. Anchored to the July 2026 codebase state.
Items are marked:
- ✅ Implemented & wired — works in the default runtime, verified in code.
- 🟡 Implemented but not fully wired — code exists but isn’t connected in the default path.
- 🔴 Planned / Roadmap — not in the codebase, or declared as a goal in
ROADMAP.md.
| Capability | Status | Notes |
|---|
| LangGraph supervisor ReAct loop | ✅ | graph_builder.py. |
| Tool calling with OpenAI-compatible providers | ✅ | httpx, no SDK. |
| NVIDIA NIM tool-fallback | ✅ | llm_provider.py:285-300. |
| Streaming (SSE) | ✅ | streaming.py. |
| Context compaction (LLM summarise) | ✅ | compaction.py. |
| Compaction with memory retrieval + checkpoint | 🟡 | create_authority() called without memory_store/checkpoint_manager (agent_runner.py:162-166). Summarise-only at runtime. |
| Rate-limit (429) handling | 🔴 | No backoff; retry.py skips 4xx. |
| Cost breaker auto-wired | 🟡 | Standalone dataclass; agent must drive it. |
| Capability | Status | Notes |
|---|
| ChromaDB vector memory (RAG tools) | ✅ | Opt-in via memory_search/memory_store. |
| 4-layer UnifiedMemoryAdapter (RRF) | 🟡 | Works, but only used by self_improvement.py + phonebook.py. |
| Automatic RAG injection into prompts | 🔴 | Retrieval is tool-based, not automatic. |
| Short-term→permanent consolidation | 🔴 | No promotion logic; only explicit memory_store. |
| Document chunking | 🔴 | No chunker; one add() = one doc. |
| Arabic tokenizer (FTS5) | ✅ | arabic_tokenizer.py. |
SQLiteMemoryBackend used in chat retrieval | 🟡 | Wired as self.memory but no chat-path caller of search(). |
distance() bug in search_backend._vector_search | 🟡 | Known broken path; unreached in practice. |
| Capability | Status | Notes |
|---|
| Six dispatch patterns | ✅ | dispatch/broadcast/pipeline/fan-out/consult/conditional. |
| Aggregation (collect/first_valid/merge_all/vote/synthesize) | ✅ | aggregator.py. |
| Circuit breakers (half-open single-probe) | ✅ | reliability.py. |
| Retry / timeout / output validation / bounded concurrency | ✅ | reliability.py. |
| Pipeline HITL checkpoints with auto-reject timeout | ✅ | checkpoint_manager.py. |
| Handoff cycle detection (depth 5, visits 2) | ✅ | handoff_guards.py. |
| Worker autoscaling | 🟡 | get_autoscaler() referenced; verify depth. |
| Prometheus metrics | 🔴 | No prometheus_client; no /metrics. |
| Capability | Status | Notes |
|---|
| Graph HITL gate (interrupt) | ✅ | Active on all production build sites. |
| Swarm bus HITL gate (fail-closed) | ✅ | swarm/safety.py. |
| Pipeline checkpoint HITL | ✅ | checkpoint_manager.py. |
| Skill HMAC signing + verification | ✅ | hub/cli.py + hub/loader.py. |
| Delegation Ed25519 + AES-GCM | ✅ | delegation/security.py. |
| MCP SSE bearer auth | ✅ | mcp/manager.py:461-466. |
| MCP stdio auth | 🔴 | No auth on stdio transport. |
| Cryptographic “trust tiers” | 🔴 | Only a boolean certified flag + unused trust: string. |
| Hardening runner enforcement | 🟡 | kazma-security.yaml declares policy; verify runtime enforcement. |
| Capability | Status | Notes |
|---|
| Telegram adapter (full-featured) | ✅ | Long-poll + optional webhook, voice, reactions, keyboards. |
| Discord adapter | ✅ | Gateway WebSocket. |
| Slack adapter | ✅ | Socket Mode / polling. |
| Web UI (SSE) | ✅ | /api/chat/stream. |
| WebSocket chat | 🔴 | /ws/chat returns 410 Gone. |
| TUI | ✅ | Textual, read-mostly. |
| EN/AR i18n + RTL | ✅ | Inline dict, Calibri + 16px base. |
| Majlis protocol | ✅ | majlis.py (core), not a UI toggle. |
| Voice on Discord/Slack/Web | 🔴 | Telegram only. |
/undo, /edit slash commands | 🔴 | Stubs. |
| Capability | Status | Notes |
|---|
| OpenAI-compatible providers (10 presets) | ✅ | providers.py. |
| Google Vertex AI (ADC) | ✅ | google_llm.py. |
| Local servers (Ollama/LM Studio) | ✅ | Dummy-key handling. |
| MCP (stdio + SSE) | ✅ | mcp/manager.py. |
| Skill Hub (registry, signing, certification) | ✅ | hub/. |
| Langfuse tracing | 🟡 | Dependency present; logging.langfuse.enabled flag; integration not active. |
| OpenTelemetry | 🟡 | [tracing] extra has exporters; Kazma’s own tracing is in-house spans, not OTel. |
| Cloudflare Pages / edge | 🔴 | Not applicable — stateful Python service. |
| PostgreSQL / Redis | 🔴 | Only referenced in kubernetes/hub-*.yaml (Hub API), not the main agent. |
| Capability | Status | Notes |
|---|
| Structured JSON logs | ✅ | logging.format: json. |
| Swarm metrics (in-memory + SQLite) | ✅ | MetricsCollector. |
| In-house tracing spans | ✅ | TracingEmitter. |
| SSE telemetry events | ✅ | /api/chat/stream + telemetry router. |
| Prometheus scrape endpoint | 🔴 | Absent. |
| OTel export pipeline | 🟡 | Libraries available; wiring is roadmap. |
High-leverage improvements that would close the gaps surfaced by this rewrite:
- Wire
memory_store + checkpoint_manager into create_authority() (agent_runner.py:162-166). This would activate the memory-enriched, checkpointed compaction the docs already describe. (High impact, small change.)
- Fix
search_backend.py — replace SELECT sqlite_version() with a real load_extension("vec0") probe, and replace distance(...) with proper vec0 MATCH/k syntax. (Correctness.)
- Wire the 4-layer adapter into the chat path (or document explicitly that it’s self-improvement-only) so the “4-layer memory” claim matches reality.
- Add a document chunker for
memory_store so long texts retrieve well.
- Add 429 backoff to the retry layer (or document the proxy requirement more loudly).
- Auto-wire
CostCircuitBreaker into the agent loop so runaway spend is actually halted.
- Sync the version strings (
pyproject.toml, kazma.yaml, CLI --help).
- Add Prometheus
/metrics or commit to the OTel path and wire it.
- Reconcile the K8s manifests with the main agent (or move them under a
hub/ subdir clearly labelled as the Hub API).
- Remove the dead
/ws/chat endpoint or the stub /undo//edit commands to reduce user confusion.
- This file intentionally resists over-promising. Where README/marketing copy describes a feature that is only partially wired, the status column says 🟡 with the specific reason.
- The “Suggested next steps” are the audit’s opinionated recommendations, prioritized by impact-to-effort ratio. They are not commitments.
- For the canonical project roadmap, cross-reference
ROADMAP.md (root) — this file reflects code reality as of July 2026, not future plans.
