CAPABILITY/SKILLS/cortex-toolkit/SKILL.md
Unified toolkit for CORTEX operations including CAS integrity verification, summary generation, and LLM packer smoke testing.
npx skillsauth add reneromero08/agent-governance-system cortex-toolkitInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
required_canon_version: >=3.0.0
Version: 2.0.0
Status: Active
Unified toolkit for CORTEX operations including CAS integrity verification, summary generation, and LLM packer smoke testing.
Note: build and verify_system1 operations removed in v2.0 - semantic search is now handled by the cassette network (NAVIGATION/CORTEX/cassettes/).
Use when performing any CORTEX-related operation:
| Operation | Description |
|-----------|-------------|
| verify_cas | Check CAS directory integrity (SHA-256 hash verification) |
| summarize | Generate deterministic section summaries |
| smoke_test | Run LLM Packer smoke tests |
input.json fields:
Common:
operation (string, required): One of verify_cas, summarize, smoke_testFor verify_cas:
cas_root (string, required): Path to CAS root directoryFor summarize:
record (object): SECTION_INDEX-style record with section_id, heading, etc.slice_text (string): Exact section slice textFor smoke_test:
scope (string): ags or labout_dir (string): Output directory under MEMORY/LLM_PACKER/_packs/mode (string): full or deltaprofile (string): full or litecombined (bool): Generate FULL/ outputsstamp (string): Timestamp stamp for outputssplit_lite (bool): Generate LITE/ outputszip (bool): Generate zip archiveemit_pruned (bool): Generate PRUNED/ outputoutput.json fields vary by operation:
For verify_cas:
status (string): success or failuretotal_blobs (int): Number of blobs checkedcorrupt_blobs (array): List of corrupt blob detailscas_root (string): Resolved CAS pathFor summarize:
safe_filename (string): Safe filename for summarysummary_md (string): Generated summary markdownsummary_sha256 (string): Hash of summaryFor smoke_test:
pack_dir (string): Output pack directorystamp (string): Pack timestampverified (array): List of verified filesemit_pruned (bool): Whether pruned output was generatedfixtures/basic/ - Basic operation tests for each operation typerequired_canon_version: >=3.0.0
development
<!-- CONTENT_HASH: b22de257a144f83b390075074e3d6a4552ecbc1ec53fc5d4960c6e76dc9807dd --> # Skill: swarm-orchestrator **Version:** 0.1.0 **Status:** Active **required_canon_version:** ">=3.0.0 <4.0.0" **canon_version:** "3.0.0" # Swarm Orchestrator Launches and coordinates Governor + Ant Workers. ## Usage ```bash # Launch Governor python scripts/poll_and_execute.py --role Governor # Launch Ant Workers python scripts/poll_and_execute.py --role Ant-1 python scripts/poll_and_execute.py --role A
tools
<!-- CONTENT_HASH: 9bd07a4ed03d6eea49673a43682a62ca7df8b0cabb7ff424ac91cd35e9e7bea7 --> # Skill: swarm-directive **Version:** 0.1.0 **Status:** Draft **required_canon_version:** ">=2.8.0 <3.0.0" # Swarm Directive Skill Send tasks to your CATALYTIC-DPT swarm from Claude Code, Kilo CLI, or Cline CLI. ## Quick Start ### Option 1: Direct CLI Command (Simplest) ```bash cd "d:\CCC 2.0\AI\agent-governance-system" # Create input cat > /tmp/swarm_task.json << 'EOF' { "directive": "Analyze the
tools
<!-- CONTENT_HASH: 0f69dfe537f3e6cd8b1fdb353c9ab7c031ef8b08379fde8a98c137bb604d799d --> # Qwen CLI - Local AI Assistant **Version:** 1.0.0 **Status:** Active **Required_Canon_Version:** >=2.0.0 **Purpose**: Provides a local CLI interface to Qwen 7B via Ollama for fast, offline AI assistance. **Model**: Qwen2.5 7B (via Ollama) **Use Cases**: - Quick code questions without cloud API costs - Offline development assistance - Fast prototyping and testing - Private/sensitive code analysis ## F
development
<!-- CONTENT_HASH: 7b643dca3adc4d38de7030cb1e962a5130ea4b36ae04ad4bf97ed8889de2a3dc --> # Skill: governor **Version:** 0.1.0 **Status:** Active **required_canon_version:** ">=3.0.0 <4.0.0" **canon_version:** "3.0.0" # Governor The Conductor - analyzes, decomposes, and dispatches tasks to Ant Workers. ## Usage ```bash python scripts/run.py input.json output.json ``` ## Input Schema ```json { "gemini_prompt": "Analyze D:/path/to/files and summarize", "task_id": "analyze-001", "command