{
  "markdown": "# mcp-cognitive-substrate\n\nmcp-name: io.github.JaysonAIOnline/cognitive-substrate\n\nA **28-layer cognitive substrate** with cross-session **Tree-of-Thoughts (ToT) evolutionary memory**, conditional self-telemetry, and **A2A tools** for MCP agents.\n\nLets agents reason through a validated 28-layer substrate, evolve memory across sessions, and communicate with peer agents — all in one pip-installable package.\n\n## Features\n\n- **28-layer cognitive substrate with Pydantic validation** — `CognitiveSubstrate` validates reasoning through 6 families / 28 layers.\n- **Cross-session ToT evolutionary memory** — SQLite-backed tree-of-thoughts nodes + substrate history; pruned branches become lessons for future sessions.\n- **Robust stack-based JSON parser** — no regex; handles nested brackets, escaped strings, embedded code fences (`robust_slice` / `robust_json_slice`).\n- **Self-telemetry tool** — `get_cognitive_tree_state` returns active paths and pruned branches for a session.\n- **Post-execution storage loop** — `store_5key_telemetry` auto-saves compact 5-key telemetry (foundations, metacognition, defensive, resource, utility).\n- **7 reasoning paradigms** — deductive, inductive, abductive, analogical, causal, syllogistic, falsification.\n- **A2A tools** — list, discover, call, and orchestrate peer agents.\n- **MCP server** — exposes everything as tools via the `cognitive-substrate` CLI.\n\n## Install\n\n```bash\npip install mcp-cognitive-substrate\n```\n\nOr install from source:\n\n```bash\ngit clone https://github.com/JaysonAIOnline/mcp-cognitive-substrate.git\ncd mcp-cognitive-substrate\npip install -e .[test]\n```\n\nRequires Python **>= 3.11**.\n\n## Quick Start\n\n```python\nfrom mcp_cognitive_substrate.substrate import CognitiveSubstrate\nfrom mcp_cognitive_substrate.memory import get_cognitive_tree_state, store_5key_telemetry\n\nsubstrate = CognitiveSubstrate()\nresponse = substrate.run(\"Your user prompt here\")\nprint(response[\"layers_applied\"], \"layers applied\")\nprint(response[\"substrate_verdict\"])\n```\n\n## Usage\n\n### 28-layer substrate\n\n```python\nfrom mcp_cognitive_substrate import substrate\n\n# Layer count and schema\nprint(substrate.layer_count())        # 28\nprint(substrate.SUBSTRATE_SCHEMA)     # the full 6-family schema\n\n# Validate a prompt through the substrate\nresult = substrate.CognitiveSubstrate(session_id=\"s1\").run(\"deploy safely\")\nprint(result[\"substrate_verdict\"])    # heuristic pruning verdict\n\n# Run a single paradigm\nfrom mcp_cognitive_substrate import run_paradigm\nprint(run_paradigm(\"14_idempotency_side_effect_audit\", {\"evaluate_branch\": True}))\n```\n\n### Cross-session ToT evolutionary memory\n\n```python\nfrom mcp_cognitive_substrate.memory import (\n    store_5key_telemetry,\n    get_cognitive_tree_state,\n    prune_failed_approach,\n)\n\nnode_id = store_5key_telemetry(\n    session_id=\"session-a\",\n    payload={\n        \"foundations\": {\"premise_validation\": \"assuming deps\", \"state_hash\": \"h\", \"falsification_notes\": \"deps missing\"},\n        \"defensive\": {\"blast_radius\": \"unpredictable\", \"is_idempotent\": True, \"invariant_rule\": \"r\"},\n        \"resource\": {\"big_o\": \"o(n)\", \"latency_bottleneck\": \"none\"},\n        \"utility\": {\"load_summary\": \"pin versions to deploy\", \"checklist_verified\": True},\n        \"metacognition\": {\"self_critique\": \"c\", \"drift_pct\": 0.1},\n    },\n    score_delta=-110.0,\n)\nprune_failed_approach(node_id)\nstate = get_cognitive_tree_state(\"session-a\", include_pruned=True)\nprint(state[\"active_path_count\"], state[\"pruned_branch_count\"])\n```\n\n### Stack-based JSON parser\n\n```python\nfrom mcp_cognitive_substrate.memory import robust_slice, robust_json_slice\n\ncleaned, payload = robust_slice('prefix {\"a\": {\"b\": [1, 2]}, \"c\": \"x\"} suffix')\n# payload == {\"a\": {\"b\": [1, 2]}, \"c\": \"x\"}; cleaned == \"prefix suffix\"\n```\n\n### 7 reasoning paradigms + A2A\n\n```python\nfrom mcp_cognitive_substrate import reason, a2a_list, a2a_call, a2a_orchestrate\n\nprint(reason(\"Solve X\", reasoning_type=\"abductive\", depth=3)[\"steps\"])\nprint(a2a_list())\nprint(a2a_call(\"peer-agent\", \"hello\"))\nprint(a2a_orchestrate(\"hi\", capability=\"memory\"))\n```\n\n### As an MCP server\n\n```bash\ncognitive-substrate          # starts stdio MCP server\ncognitive-substrate --info   # prints package summary\n```\n\nAll of the above — substrate paradigms, extraction/evaluation, memory store/recall, ToT lessons, tree-state telemetry, JSON parsing, reasoning plans, and A2A — are exposed as MCP tools.\n\n## Testing\n\n```bash\npip install -e .[test]\npython -m pytest src/tests -q     # 16 tests\n```\n\n## License\n\n[MIT](LICENSE)",
  "bytes": 4513,
  "sha": "1f26278130c08200f3ddddeb67495235e598edd107454bbddc8f5887b0644984",
  "repo_slug": "jaysonaionline/mcp-cognitive-substrate",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/mcp_io_github_jaysonaionline_cognitive_subst_13b018ca/readme"
}