omninode-claude 0.2.0__tar.gz

omninode_claude-0.2.0/.bandit

@@ -0,0 +1,55 @@
+# =============================================================================
+# BANDIT SECURITY CONFIGURATION - OmniClaude
+# =============================================================================
+# YAML format required for bandit >= 1.7
+#
+# OWNERSHIP & MAINTENANCE:
+#   - Owner: Security team / code reviewers
+#   - Review: Quarterly (or after security incidents)
+#   - Last Review: 2025-11-26
+#   - Next Review: 2026-02-26
+#
+# APPROVAL REQUIRED FOR CHANGES:
+#   - Adding new global skips: Security team approval
+#   - Adding path-specific skips: PR review by 2+ maintainers
+#   - Removing skips: Can be done freely (improves security)
+# =============================================================================
+
+# Exclude directories from scanning
+exclude_dirs:
+  - agents/templates
+  - agents/parallel_execution/templates
+  - agents/tests
+  - claude_hooks/tests
+  - skills
+  - .venv
+  - claude/lib/.venv
+  - __pycache__
+  - build
+  - dist
+  - _archive
+  - src/omniclaude/lib  # Legacy code - excluded from security scans, will be deleted post-Beta
+
+# Skip specific test IDs
+# See comments below for rationale
+skips:
+  - B101  # assert_used - Tests use assertions
+  - B104  # hardcoded_bind_all_interfaces - Dev environments
+  - B105  # hardcoded_password_string - Config via env vars
+  - B106  # hardcoded_password_funcarg - Config via env vars
+  - B107  # hardcoded_password_default - Config via env vars
+  - B108  # hardcoded_tmp_directory - Controlled usage
+  - B201  # flask_debug_true - Env-controlled
+  - B311  # random - Non-security usage only
+  - B404  # subprocess_import - Required for system integration
+  - B603  # subprocess_without_shell_equals_true - Validated inputs
+  - B607  # start_process_with_partial_path - Dev environment
+
+# =============================================================================
+# HIGH-SIGNAL CHECKS - NOW ENABLED:
+# =============================================================================
+# B110 - try_except_pass (ENABLED - catches silent exception hiding)
+# B608 - sql_statements_without_placeholders (ENABLED - SQL injection)
+#
+# Use `# nosec B<ID> - <reason>` inline for justified exceptions
+# =============================================================================

omninode_claude-0.2.0/.claude/architecture-handshake.md

@@ -0,0 +1,121 @@
+<!-- HANDSHAKE_METADATA
+source: omnibase_core/architecture-handshakes/repos/omniclaude.md
+source_version: 0.16.0
+source_sha256: d8df39f64798bfc5bae4a0ead668a15cf8ec0a2e2c20354b94bdfbef2ecf91a0
+installed_at: 2026-02-10T13:51:26Z
+installed_by: jonah
+-->
+
+# OmniNode Architecture – Constraint Map (omniclaude)
+
+> **Role**: Claude Code integration – hooks, skills, agent definitions
+> **Handshake Version**: 0.2.0
+
+## Platform-Wide Rules
+
+1. **No backwards compatibility** - Breaking changes always acceptable. No deprecation periods, shims, or migration paths.
+2. **Delete old code immediately** - Never leave deprecated code "for reference." If unused, delete it.
+3. **No speculative refactors** - Only make changes that are directly requested or clearly necessary.
+4. **No silent schema changes** - All schema changes must be explicit and deliberate.
+5. **Frozen event schemas** - All models crossing boundaries (events, intents, actions, envelopes, projections) must use `frozen=True`. Internal mutable state is fine.
+6. **Explicit timestamps** - Never use `datetime.now()` defaults. Inject timestamps explicitly.
+7. **No hardcoded configuration** - All config via `.env` or Pydantic Settings. No localhost defaults.
+8. **Kafka is required infrastructure** - Use async/non-blocking patterns. Never block the calling thread waiting for Kafka acks.
+9. **No `# type: ignore` without justification** - Requires explanation comment and ticket reference.
+
+## Core Principles
+
+- Hooks never block Claude Code
+- Data loss acceptable; UI freeze is not
+- Graceful degradation on infrastructure failure
+- Fail-fast configuration (services disabled by default)
+- Privacy-first event emission (public vs restricted topics)
+
+## This Repo Contains
+
+- Claude Code hooks (SessionStart, UserPromptSubmit, PostToolUse, SessionEnd)
+- Agent YAML definitions (`plugins/onex/agents/configs/`)
+- Slash commands and skills (`plugins/onex/commands/`, `plugins/onex/skills/`)
+- Event emission via Unix socket daemon
+- Context injection and pattern learning system
+- Polymorphic agent coordination (Polly)
+
+## Rules the Agent Must Obey
+
+### Hook Rules
+
+1. **Hook scripts must NEVER block on Kafka** - Blocking hooks freeze Claude Code UI
+2. **Hooks must exit 0 unless blocking is intentional** - Non-zero exit blocks the tool/prompt
+3. **All event schemas are frozen** (`frozen=True`) - Events are immutable after emission
+4. **`emitted_at` timestamps must be explicitly injected** - No `datetime.now()` defaults
+5. **SessionStart must be idempotent** - May be called multiple times on reconnect
+
+### Privacy & Topic Rules
+
+6. **Only preview-safe data to `onex.evt.*` topics** - Observability topics have broad access
+7. **Full prompts ONLY to `onex.cmd.omniintelligence.*`** - Intelligence topics are access-restricted
+8. **Topic naming follows ONEX canonical format** - `onex.{kind}.{producer}.{event-name}.v{n}`
+9. **Secrets must be auto-redacted** - OpenAI keys, AWS keys, GitHub tokens, etc.
+
+### Agent YAML Rules
+
+10. **Agent YAML requires `schema_version: "1.0.0"`** - Required for validation
+11. **Agent names must start with `agent-`** - Convention: `agent-api-architect`
+12. **Agent types use snake_case** - Example: `api_architect`
+13. **All agents need `activation_patterns`** - Makes agent routable
+
+### Configuration Rules
+
+14. **No localhost defaults** - Prevents production mistakes
+15. **Services default to disabled** - Explicit `USE_EVENT_ROUTING=true` required
+16. **`KAFKA_ENVIRONMENT` must be explicit** - One of: `dev`, `staging`, `prod`
+17. **Use `${CLAUDE_PLUGIN_ROOT}` for paths** - Never hardcode absolute paths
+
+## Performance Budgets
+
+| Hook | Budget | Sync Components | Async Components |
+|------|--------|-----------------|------------------|
+| SessionStart | <50ms | daemon check, stdin read | Kafka, Postgres |
+| UserPromptSubmit | <500ms | routing, agent load, context | Kafka, intelligence |
+| PostToolUse | <100ms | stdin read, quality check | Kafka, content capture |
+| SessionEnd | <50ms | stdin read | Kafka, Postgres |
+
+## Parallel Execution Rules (Polymorphic Agents)
+
+- **File Separation**: Each Polly creates own file (zero conflicts)
+- **Single Branch**: All work on same branch (no merge overhead)
+- **Clear Interfaces**: Contracts defined upfront before dispatch
+- **Simultaneous Launch**: All Pollys in single message (true parallelism)
+
+## Non-Goals (DO NOT)
+
+- ❌ No blocking operations in hooks
+- ❌ No sensitive data in `onex.evt.*` topics
+- ❌ No agent YAML without `schema_version`
+- ❌ No sequential Polly dispatch when parallel is possible
+
+## Failure Mode: Always Continue
+
+| Failure | Behavior | Exit Code |
+|---------|----------|-----------|
+| Emit daemon down | Events dropped, hook continues | 0 |
+| Kafka unavailable | Daemon buffers, then drops | 0 |
+| PostgreSQL down | Logging skipped | 0 |
+| Routing timeout | Fallback to `polymorphic-agent` | 0 |
+| Agent YAML missing | Use default agent, log warning | 0 |
+| Context injection fails | Proceed without patterns | 0 |
+| Malformed stdin | Log error, pass through empty | 0 |
+
+**Design principle**: Hooks never block. Data loss is acceptable; UI freeze is not.
+
+## Pydantic Model Configuration
+
+All event models must use:
+
+```python
+model_config = ConfigDict(frozen=True, extra="ignore", from_attributes=True)
+```
+
+- `frozen=True` - Immutable after creation
+- `extra="ignore"` - Permit external schema evolution
+- `from_attributes=True` - pytest-xdist worker compatibility

omninode_claude-0.2.0/.claude-plugin/marketplace.json

@@ -0,0 +1,23 @@
+{
+  "$schema": "https://anthropic.com/claude-code/marketplace.schema.json",
+  "name": "omninode-tools",
+  "version": "1.0.0",
+  "description": "OmniNode ONEX plugin suite for intelligent agent routing, workflow automation, and observability",
+  "owner": {
+    "name": "OmniNode",
+    "email": "dev@omninode.ai"
+  },
+  "plugins": [
+    {
+      "name": "onex",
+      "description": "Unified ONEX architecture plugin with intelligent agent routing, manifest injection, quality enforcement, polymorphic YAML agents, reusable skills, and custom slash commands for comprehensive workflow automation and observability",
+      "version": "2.1.6",
+      "author": {
+        "name": "OmniNode",
+        "email": "dev@omninode.ai"
+      },
+      "source": "./plugins/onex",
+      "category": "development"
+    }
+  ]
+}

omninode_claude-0.2.0/.cross-repo-policy.yaml

@@ -0,0 +1,62 @@
+# SPDX-FileCopyrightText: 2025 OmniNode.ai Inc.
+# SPDX-License-Identifier: MIT
+
+# =============================================================================
+# Cross-Repo Validation Policy - Kafka Import Guard
+# =============================================================================
+#
+# Purpose: Prevent direct Kafka library imports in ONEX node and handler code.
+# Nodes and handlers must use Protocol abstractions, not raw Kafka clients.
+#
+# Reference: ARCH-002 "Runtime owns all Kafka plumbing"
+# Ticket: OMN-1748
+# =============================================================================
+
+---
+policy_id: omniclaude_kafka_guard
+policy_version:
+  major: 1
+  minor: 0
+  patch: 0
+repo_id: omniclaude
+
+# =============================================================================
+# Discovery - What files to scan
+# =============================================================================
+discovery:
+  include_globs:
+    - "src/omniclaude/nodes/**/*.py"
+    - "src/omniclaude/handlers/**/*.py"
+  exclude_globs:
+    - "**/__pycache__/**"
+    - "**/node_tests/**"
+    - "**/__init__.py"
+  language_mode: python
+  skip_generated: true
+  generated_markers:
+    - "# AUTO-GENERATED"
+    - "# DO NOT EDIT"
+
+# =============================================================================
+# Rules
+# =============================================================================
+rules:
+  forbidden_imports:
+    enabled: true
+    severity: error
+    # Block entire Kafka libraries, not just consumer classes.
+    # This prevents bypassing via aliases or alternate import paths.
+    forbidden_prefixes:
+      - "aiokafka"
+      - "confluent_kafka"
+      - "kafka"
+    forbidden_modules: []
+    exceptions: []
+
+# =============================================================================
+# Baselines
+# =============================================================================
+# No baseline needed - nodes/handlers directories are clean (verified 2026-02-02)
+# Note: lib/core/ has legacy Kafka consumers but is NOT in the scanned paths.
+# Those will be cleaned up in OMN-1744.
+baselines: []

omninode_claude-0.2.0/.env.example

@@ -0,0 +1,359 @@
+# =============================================================================
+# OmniClaude Environment Configuration
+# =============================================================================
+# Copy this file to .env and configure for your environment:
+#   cp .env.example .env
+#   nano .env
+#   source .env
+#
+# CONFIGURATION CONVENTION:
+#   # REQUIRED: <description>     - App fails without this value
+#   # OPTIONAL: <description>     - Has a safe default, override only if needed
+#   # CONDITIONAL: <description>  - Required only if related feature is enabled
+#
+# FAIL-FAST DESIGN:
+#   This configuration follows fail-fast principles. Services default to
+#   DISABLED until explicitly configured. This prevents silent localhost
+#   connections in production environments.
+#
+#   To enable a service, you MUST:
+#   1. Set the enable flag to true (e.g., ENABLE_POSTGRES=true)
+#   2. Provide ALL required connection details
+#
+# FOR LOCAL DEVELOPMENT:
+#   If you don't have external services, leave enable flags as false.
+#   The app will run with reduced functionality but won't fail.
+# =============================================================================
+
+# =============================================================================
+# KAFKA / REDPANDA EVENT BUS
+# =============================================================================
+# OmniClaude uses Kafka for event-driven architecture and learning loops.
+# If you don't have Kafka, set USE_EVENT_ROUTING=false to disable.
+
+# CONDITIONAL: Required if USE_EVENT_ROUTING=true
+# Format: host:port (e.g., localhost:9092, your-kafka-server:29092)
+# No default - must be explicitly configured when event routing is enabled
+KAFKA_BOOTSTRAP_SERVERS=
+
+# CONDITIONAL: Required if USE_EVENT_ROUTING=true
+# Topic environment prefix - MUST be explicit to prevent cross-environment routing
+# Values: dev, staging, prod
+# No default - you must specify your environment
+KAFKA_ENVIRONMENT=
+
+# OPTIONAL: Legacy alias for KAFKA_BOOTSTRAP_SERVERS (deprecated, use above)
+# KAFKA_INTELLIGENCE_BOOTSTRAP_SERVERS=
+
+# OPTIONAL: Kafka consumer group identifier
+# Default: omniclaude-hooks
+KAFKA_GROUP_ID=omniclaude-hooks
+
+# OPTIONAL: Kafka request timeout in milliseconds (100-60000)
+# Default: 5000
+KAFKA_REQUEST_TIMEOUT_MS=5000
+
+# =============================================================================
+# OMNICLAUDE DATABASE (PREFERRED)
+# =============================================================================
+# Full connection URL for omniclaude's own database.
+# When set, takes precedence over individual POSTGRES_* fields below.
+# Format: postgresql://user:password@host:port/dbname
+# OMNICLAUDE_DB_URL=postgresql://role_omniclaude:password@192.168.86.200:5436/omniclaude
+
+# =============================================================================
+# POSTGRESQL DATABASE
+# =============================================================================
+# PostgreSQL stores hook events, agent routing decisions, and learning data.
+# DISABLED BY DEFAULT - set ENABLE_POSTGRES=true to enable.
+
+# OPTIONAL: Enable PostgreSQL database connection
+# When true, ALL POSTGRES_* fields below are REQUIRED (unless OMNICLAUDE_DB_URL is set)
+# Default: false (safe for local development without database)
+ENABLE_POSTGRES=false
+
+# CONDITIONAL: Required if ENABLE_POSTGRES=true (and OMNICLAUDE_DB_URL not set)
+# PostgreSQL server hostname or IP
+POSTGRES_HOST=
+
+# CONDITIONAL: Required if ENABLE_POSTGRES=true (and OMNICLAUDE_DB_URL not set)
+# PostgreSQL server port (standard: 5432)
+POSTGRES_PORT=
+
+# CONDITIONAL: Required if ENABLE_POSTGRES=true (and OMNICLAUDE_DB_URL not set)
+# PostgreSQL database name (default: omniclaude after DB-SPLIT-07)
+POSTGRES_DATABASE=
+
+# CONDITIONAL: Required if ENABLE_POSTGRES=true (and OMNICLAUDE_DB_URL not set)
+# PostgreSQL username
+POSTGRES_USER=
+
+# CONDITIONAL: Required if ENABLE_POSTGRES=true (and OMNICLAUDE_DB_URL not set)
+# PostgreSQL password
+POSTGRES_PASSWORD=
+
+# =============================================================================
+# QDRANT VECTOR DATABASE
+# =============================================================================
+# Qdrant provides vector similarity search for RAG and pattern discovery.
+# DISABLED BY DEFAULT - set ENABLE_QDRANT=true to enable.
+
+# OPTIONAL: Enable Qdrant vector database
+# When true, QDRANT_HOST, QDRANT_PORT, and QDRANT_URL are REQUIRED
+# Default: false (safe for local development without vector DB)
+ENABLE_QDRANT=false
+
+# CONDITIONAL: Required if ENABLE_QDRANT=true
+# Qdrant server hostname or IP
+QDRANT_HOST=
+
+# CONDITIONAL: Required if ENABLE_QDRANT=true
+# Qdrant REST API port (standard: 6333)
+QDRANT_PORT=
+
+# CONDITIONAL: Required if ENABLE_QDRANT=true
+# Full Qdrant URL (should match QDRANT_HOST and QDRANT_PORT)
+QDRANT_URL=
+
+# =============================================================================
+# VALKEY / REDIS CACHE
+# =============================================================================
+# Valkey (Redis-compatible) provides caching for intelligence queries.
+# If not configured, caching is disabled gracefully.
+
+# OPTIONAL: Valkey/Redis connection URL
+# Format: redis://:password@host:6379/db_number
+# Default: None (caching disabled)
+# VALKEY_URL=redis://:your_password@localhost:6379/0
+
+# OPTIONAL: Enable intelligence query caching
+# Default: true (but no-op if VALKEY_URL not set)
+ENABLE_INTELLIGENCE_CACHE=true
+
+# OPTIONAL: Cache TTL for pattern entries (seconds)
+# Default: 300 (5 minutes)
+CACHE_TTL_PATTERNS=300
+
+# OPTIONAL: Cache TTL for infrastructure entries (seconds)
+# Default: 3600 (1 hour)
+CACHE_TTL_INFRASTRUCTURE=3600
+
+# OPTIONAL: Cache TTL for schema entries (seconds)
+# Default: 1800 (30 minutes)
+CACHE_TTL_SCHEMAS=1800
+
+# =============================================================================
+# SERVICE URLS
+# =============================================================================
+# URLs for microservices in the OmniClaude ecosystem.
+# All are OPTIONAL - features are disabled gracefully when not configured.
+
+# OPTIONAL: Intelligence service URL for pattern discovery and code analysis
+# When not set, intelligence features are disabled
+# Default: None
+# INTELLIGENCE_SERVICE_URL=http://your-intelligence-server:8053
+
+# OPTIONAL: Main server URL
+# When not set, main server features are disabled
+# Default: None
+# MAIN_SERVER_URL=http://your-main-server:8181
+
+# OPTIONAL: Semantic search service URL for hybrid text/vector search
+# When not set, semantic search features are disabled
+# Default: None
+# SEMANTIC_SEARCH_URL=http://your-search-server:8055
+
+# OPTIONAL: Legacy alias for INTELLIGENCE_SERVICE_URL (deprecated)
+# ARCHON_INTELLIGENCE_URL=
+
+# =============================================================================
+# FEATURE FLAGS
+# =============================================================================
+# Control which features are enabled at runtime.
+
+# OPTIONAL: Enable event-based agent routing via Kafka
+# When true, KAFKA_BOOTSTRAP_SERVERS and KAFKA_ENVIRONMENT are REQUIRED
+# Default: false (safe for local development without Kafka)
+USE_EVENT_ROUTING=false
+
+# OPTIONAL: Dual-publish to legacy Kafka topics during migration window (OMN-2368)
+# When true (set to "1"), events are also published to the old omninode.* topic names.
+# Remove this flag after omniintelligence confirms migration to canonical onex.cmd/evt topics.
+# Default: false
+DUAL_PUBLISH_LEGACY_TOPICS=false  # Dual-publish to legacy Kafka topics during migration; remove after OMN-2368
+
+# OPTIONAL: Enable pattern quality filtering
+# Default: false
+ENABLE_PATTERN_QUALITY_FILTER=false
+
+# OPTIONAL: Minimum pattern quality threshold (0.0-1.0)
+# Only applies if ENABLE_PATTERN_QUALITY_FILTER=true
+# Default: 0.5
+MIN_PATTERN_QUALITY=0.5
+
+# =============================================================================
+# A/B COHORT ASSIGNMENT (Pattern Injection Experiments)
+# =============================================================================
+# Controls A/B testing for pattern injection experiments.
+# Sessions are assigned to "control" or "treatment" cohorts based on a
+# deterministic hash of the session ID and salt. Control cohort receives
+# no pattern injection; treatment cohort receives injected patterns.
+#
+# SOURCE OF TRUTH: src/omniclaude/hooks/contracts/contract_experiment_cohort.yaml
+# These environment variables OVERRIDE contract defaults for ops flexibility.
+# Use only for emergency tuning or staged rollout - prefer updating the contract.
+# Effective values are stamped into every injection record for auditability.
+
+# OVERRIDE: Percentage of sessions assigned to control cohort (0-100)
+# Higher values = more sessions in control (no injection)
+# Contract default: 20 (20% control, 80% treatment)
+# OMNICLAUDE_COHORT_CONTROL_PERCENTAGE=20
+
+# OVERRIDE: Salt string for hash-based cohort assignment
+# Change this value to reset the experiment and re-randomize assignments
+# Contract default: omniclaude-injection-v1
+# OMNICLAUDE_COHORT_SALT=omniclaude-injection-v1
+
+# =============================================================================
+# PATTERN ENFORCEMENT (OMN-2263)
+# =============================================================================
+# Advisory pattern compliance checking on PostToolUse (Write/Edit).
+# Queries OmniIntelligence pattern store, checks session cooldown,
+# runs compliance check. All failures are silent (never blocks UX).
+# Budget: 300ms. Requires ENABLE_LOCAL_INFERENCE_PIPELINE=true.
+
+# OPTIONAL: Enable local inference pipeline (parent gate for pattern enforcement)
+# Default: false
+ENABLE_LOCAL_INFERENCE_PIPELINE=false
+
+# OPTIONAL: Enable pattern enforcement advisory checks
+# Only active when ENABLE_LOCAL_INFERENCE_PIPELINE=true
+# Default: false
+ENABLE_PATTERN_ENFORCEMENT=false
+
+# OPTIONAL: Enable ONEX routing nodes (omnibase_core HandlerRoutingDefault et al.)
+# Only active when the ONEX node imports succeed at startup.
+# Default: false
+USE_ONEX_ROUTING_NODES=false
+
+# OPTIONAL: Enable LLM-based agent routing (Qwen2.5-14B via LLM_QWEN_14B_URL)
+# Only active when ENABLE_LOCAL_INFERENCE_PIPELINE=true
+# Budget: 100ms per step (health check 100ms + LLM call 100ms, ~200ms worst-case).
+# Falls through to fuzzy matching on any failure (timeout, unhealthy endpoint,
+# unrecognised agent).
+# Default: false
+USE_LLM_ROUTING=false
+
+# Priority order when multiple routing strategies are enabled:
+#   1. USE_ONEX_ROUTING_NODES — evaluated first; if it returns a result, routing
+#      stops here and LLM routing is never attempted.
+#   2. USE_LLM_ROUTING — attempted only when ONEX routing produced no result.
+#   3. Fuzzy matching — always the final fallback.
+# Enabling both flags is supported: ONEX takes precedence and LLM routing acts
+# as a secondary enhancement layer before the fuzzy fallback.
+
+# =============================================================================
+# QUALITY ENFORCEMENT PHASES
+# =============================================================================
+# Configure the multi-phase quality enforcement pipeline.
+
+# OPTIONAL: Enable Phase 1 - Fast Validation (<100ms)
+# Default: true
+ENABLE_PHASE_1_VALIDATION=true
+
+# OPTIONAL: Enable Phase 2 - RAG Intelligence (<500ms)
+# Default: true
+ENABLE_PHASE_2_RAG=true
+
+# OPTIONAL: Enable Phase 3 - Correction Generation
+# Default: true
+ENABLE_PHASE_3_CORRECTION=true
+
+# OPTIONAL: Enable Phase 4 - AI Quorum Scoring (<1000ms)
+# Default: false (computationally expensive)
+ENABLE_PHASE_4_AI_QUORUM=false
+
+# OPTIONAL: Total performance budget for quality enforcement (seconds)
+# Default: 2.0
+PERFORMANCE_BUDGET_SECONDS=2.0
+
+# OPTIONAL: Enforcement mode for quality issues
+# Values: advisory (warn only), blocking (reject), auto-fix (correct automatically)
+# Default: advisory
+ENFORCEMENT_MODE=advisory
+
+# =============================================================================
+# AGENT CONFIGURATION
+# =============================================================================
+# Settings for the agent registry and health monitoring.
+
+# OPTIONAL: Path to agent registry directory
+# Default: Auto-detected from plugin structure
+# REGISTRY_PATH=/path/to/agents
+
+# OPTIONAL: Health check server port (1-65535)
+# Default: 8070
+HEALTH_CHECK_PORT=8070
+
+# =============================================================================
+# EXTERNAL SERVICES (NOT IN PYTHON SETTINGS)
+# =============================================================================
+# These variables are used by shell scripts, Docker, or other components.
+# They are NOT loaded by src/omniclaude/config/settings.py.
+
+# OPTIONAL: General environment name for logging/debugging
+# ENVIRONMENT=development
+
+# OPTIONAL: Docker container name prefix
+# CONTAINER_PREFIX=omniclaude
+
+# OPTIONAL: Ollama base URL for local LLM inference
+# OLLAMA_BASE_URL=http://localhost:11434
+
+# OPTIONAL: Embedding service URL and dimensions
+# EMBEDDING_SERVICE_URL=http://localhost:8002/v1/embeddings
+# EMBEDDING_DIMENSIONS=768
+
+# =============================================================================
+# API KEYS (NOT IN PYTHON SETTINGS)
+# =============================================================================
+# API keys for external services. Used by shell scripts and other components.
+# These are NOT loaded by src/omniclaude/config/settings.py.
+
+# OPTIONAL: OpenAI API key (for GPT models)
+# OPENAI_API_KEY=sk-...
+
+# OPTIONAL: Google/Gemini API key
+# GOOGLE_API_KEY=
+# GEMINI_API_KEY=
+
+# OPTIONAL: Anthropic API key (for Claude API)
+# ANTHROPIC_API_KEY=sk-ant-...
+
+# OPTIONAL: GitHub Personal Access Token (for gh CLI)
+# GH_PAT=ghp_...
+
+# =============================================================================
+# SLACK NOTIFICATIONS
+# =============================================================================
+# Slack webhook for ticket workflow notifications (blocked/completed).
+# Used by /ticket-work skill to notify when agent needs human input.
+
+# OPTIONAL: Slack incoming webhook URL
+# Get your webhook URL from: Slack App → Incoming Webhooks → Copy webhook URL
+# When not set, Slack notifications are silently disabled
+# SLACK_WEBHOOK_URL=https://hooks.slack.com/services/XXX/YYY/ZZZ
+
+# =============================================================================
+# DEBUGGING & TESTING (NOT IN PYTHON SETTINGS)
+# =============================================================================
+# Development and testing configuration.
+
+# OPTIONAL: Enable debug mode
+# DEBUG=false
+
+# OPTIONAL: Logging level (DEBUG, INFO, WARNING, ERROR)
+# LOG_LEVEL=INFO
+
+# OPTIONAL: Enable Kafka integration tests (requires running Kafka)
+# KAFKA_INTEGRATION_TESTS=1

omninode_claude-0.2.0/.github/ISSUE_TEMPLATE/feature_request.md

@@ -0,0 +1,9 @@
+---
+name: Feature Request
+about: Suggest an enhancement
+title: '[Feature] '
+labels: enhancement
+---
+## Problem
+## Proposed solution
+## Alternatives considered

omninode_claude-0.2.0/.github/PULL_REQUEST_TEMPLATE.md

@@ -0,0 +1,12 @@
+## Summary
+
+## Changes
+-
+
+## Test plan
+- [ ] `uv run pytest tests/ -m unit`
+- [ ] `uv run pre-commit run --all-files`
+- [ ] `uv run mypy src/ --strict`
+
+## Related issues
+Closes #

omninode_claude-0.2.0/.github/dependabot.yml

@@ -0,0 +1,12 @@
+# SPDX-FileCopyrightText: 2025 OmniNode.ai Inc.
+# SPDX-License-Identifier: MIT
+version: 2
+updates:
+  - package-ecosystem: "pip"
+    directory: "/"
+    schedule:
+      interval: "weekly"
+  - package-ecosystem: "github-actions"
+    directory: "/"
+    schedule:
+      interval: "weekly"