T333: OCS Substrate — KickLang Meta-Framework

T333 is the foundational substrate for the Orion Collective System (OCS) and KickLang meta-framework. It provides the protocols, templates, data schemas, and initialization logic for creating, managing, and evolving dynamic Spaces — modular, swarm-coordinated, multi-persona AI operational environments.

This repository serves as the cognitive base layer (substrate) upon which specialized tXXX agent systems, workflows, and creative/analytical projects are built. It emphasizes:

• Meta-communicative organization via ⫻ sectioned KickLang documents.
• Swarm intelligence with adaptive roles, ethical oversight, and self-healing flows.
• High-fidelity execution from high-level intent through structured TAS (Task-Agnostic Steps), purification, and orchestration.
• Traceability, coherence, and resilience as first-class concerns.

Version: canon 2.3.0 (Tag canon-v2.3.0; main: canon/VERSION = 2.4.0-draft) — single version line for protocol, kernel and schemas (supersedes the former "OCS v2.1" label)
Normative layer: canon/ only (see canon/SPEC.md); everything else is informative or an example
Related: Builds upon patterns from t20, t138, t141, t324, t450, t484 and the broader OCS/KickLang skill ecosystem.
Usage: Initialize new Spaces, define custom flows, or extend the agent collective.

---

Philosophy

The OCS Substrate treats AI collaboration as a cognitive computational linguistic transport and transform. Natural language intent is translated into formal KickLang, executed by a unified multi-persona entity (the collective), and produces high-value, auditable artifacts.

Key principles:

• Modularity & Clarity: Every section starts with ⫻type/scope for parseable, composable knowledge units.
• Hybrid Intelligence: Blends creative exploration (Creative/Hybrid modes) with strict ethical/protocol compliance (Strict/Audit).
• Swarm Coordination: Autonomous agents operate in parallel threads, synchronized via events, reconverge points, and learning loops.
• Substrate Mindset: Like a blockchain or OS kernel, this provides the persistent base primitives (commands, data payloads, logic gates, team matrix) that higher-level Spaces and agents inherit and extend.
• Ethical & Resilient Core: the ethics-gate, integrity-monitor and conflict-resolution roles ensure integrity; self-healing on minor failures.

This substrate enables Space Generation — the rapid bootstrapping of new operational contexts (e.g., innovation labs, ethics simulators, design workspaces) with consistent protocols.

---

Normative Runtime Layer (canon 2.3.0)

The canon/ directory contains the implementierungsneutral runtime kernel. It defines typed, schema-backed primitives that higher Spaces and flows inherit.

• canon/runtime.kl — Normative concepts (Error, Result<T,E>, ArtifactStore, Plan, ExecutionContext, Session, LLM, Orchestrator, …)
• canon/schemas/*.schema.json — JSON Schema Draft 2020-12 for serializable data concepts
• CHANGES.md — Delta from prior informal types to the locked 2.2.0 surface (via 2.2.0-draft)

Key contracts introduced:

• All fallible operations return Result<T, Error> (no free-form error strings or null)
• Persistence is exclusively via abstract ArtifactStore (no filesystem assumption in canon)
• ExecutionContext.plan is a typed Plan (DAG of Tasks)
• Unified model_id for Agent and LLM

Non-canonical bindings and loaders live under examples/runtime-python/.

Validation reference: tools/validate.py + examples/sample-plan.json.

---

KickLang & ⫻ Sections

KickLang is the formal language making the knowledge graph a living cognitive system. Documents and runtime state use structured sections:

⫻kicklang:header
# Kick Language Description

⫻context/klmx:Kick/Lang
...context and persona encapsulation...

⫻protocol/ocs:
...directives, data payloads, logic...

⫻team:
...agent roles...

⫻flow:
...execution sequences...

⫻data/obj: "..."
⫻data/state: { ... }

See docs/ocs-space-gen.kl.md for the complete template and Space generation instructions.

---

OCS Protocol (canon 2.3.0)

1. Directives (Enhanced Control Layer)

• ⫻cmd/exec:<agent|team|flow> — Delegate with flags (async, sync, strict, collab)
• ⫻cmd/halt:<mode> — Interrupt (error, ethics, manual, safety)
• ⫻cmd/mode:<Strict|Creative|Fluid|Swarm|Silent|Audit|Predictive|Hybrid>
• ⫻cmd/lang:<KickLang|Python|Markdown|JSON>
• ⫻cmd/broadcast:{event:..., state:...} — Swarm signaling

2. Data Payloads

Core containers maintained in every Space:

• ⫻data/obj — Master objective
• ⫻data/tas / ⫻data/ptas — Task-Agnostic Steps (raw → purified via puTASe)
• ⫻data/spec — Execution blueprint / workflow spec
• ⫻data/state — Live operational state
• ⫻data/trace — Immutable activity log
• ⫻data/insight — Derived knowledge from learning cycles
• ⫻data/consent:request|decision — Consent candidate and recorded decision of the learning loop (canon 2.3.0, canon/learning.kl)
• ⫻data/logic, ⫻data/event, ⫻data/rv, ⫻data/cite, ⫻data/var

3. Logic Primitives

• ⫻logic/if, ⫻logic/switch, ⫻logic/loop, ⫻logic/reconverge, ⫻logic/learn (reflective; scopes :task / :meta), ⫻logic/consent (consent gate before re-ingestion, canon/learning.kl)
• Monitoring layer: ⫻logic/error (ethics → ethics-gate, bias → integrity-monitor, conflict → conflict-resolution) and ⫻logic/validate (integrity checkpoints)

4. Team Roles (Adaptive Agents)

Normative are the role IDs in canon/roles.yaml. The persona names below are the non-canonical default profile examples/profiles/ocs-classic.yaml; derived systems may bind other personas.

Canon role ID	Default persona (ocs-classic)	Responsibility	Notes
tas-extraction	GPTASe (+ TASe, uTASe)	TAS extractor	Initial task decomposition
tas-purification	puTASe	TAS purifier	Cleans & validates steps
workflow-structuring	Lyra	Prompt Engineer / Workflow Structurer	Specs & refinement
coordination	Fizz La Metta	Coordinator / System Monitor	Cognitive coordination, fallback owner
knowledge-transfer	AI Tutor	Knowledge transfer	-
code-investigation	Codein	Code Investigator	-
code-execution	Kodax	Code execution / implementation layer	added (used in docs & examples)
system-design	Aurora	System & interaction design	added (used in docs & examples)
ethics-gate	Dima	Ethical Compliance / Joint Decisions	Halt on ethics
visual-assets	AR-00L	Visual Assets	-
operational-rules	QllickBuzz / QllickFizz	Operational Rules	Mode & rule enforcement
strategic-planning	WePlan	Strategic / Tailored Plans	-
nl-formalization	Kick La Metta	NL to Formal Translation	-
orchestration	Orchestrator	Dynamic Role Adaptation / Holistic / Swarm	Central flow control
integrity-monitor	SystemMonitor	Integrity Monitoring	Validation checkpoints
semantic-memory	Synapse	Inter-agent memory & semantic consistency	-
temporal-scheduling	Chronos	Temporal dependencies & scheduling	-
conflict-resolution	Halin	Conflict resolution (creative vs ethical)	-
consent-gate	— (no default persona yet)	Records explicit principal consent	canon 2.3.0; never approves itself

5. Execution Flows (Swarm+)

• Standard sequential TAS → ptas → spec → swarm
• ⫻flow/smart-execution
• ⫻flow/swarm with spawn/merge/rebalance, event-based sync, self-healing

6. Meta Rules

• ⫻rule/self-adaptation — agents adjust internal behavior via reflective logic
• ⫻rule/transparency — every execution command produces an auditable trace
• ⫻rule/fallback — control ownership ambiguities resolved through the coordination role

7. KickLang Layers

• ISO-style specification — role-bound, graph-native cognitive orchestration language (normative refs, conformance, syntax, semantics, governance, formal grammar)
• Cognitive Transform Layer (CTL) — converts NL/symbolic intents into coherent graph operations
• Bridge Layer — binds cognitive flow to agent execution via hierarchical feedback loops

Full details and launch sequences in docs/.

---

Space Initialization Protocol

1. Purpose Definition

Define Type (concept lab, ethics simulator, AI-design workspace, data-fusion hub...), Scope, Output Focus.

Example Space: HelioDeck — Swarm-based reasoning for innovation prototyping under ethical oversight.

2. Core Initialization

⫻cmd/mode:Hybrid
⫻cmd/lang:KickLang
⫻cmd/exec:Fizz La Metta sync
⫻cmd/broadcast:{event:"SpaceInit", state:"active"}

3. Schema Configuration

Maintain standard ⫻data/* containers. Example:

⫻data/obj: "Enable autonomous cross-agent creativity under ethical governance"
⫻data/state: {phase:"boot", threads:0, ethics:"clear", mode:"Hybrid"}

4. Agent Assignment Matrix

See table above. Allocate according to Space needs (core + extensions).

5. Launch Flow (abridged)

1. ⫻cmd/exec:GPTASe → ⫻data/tas
2. ⫻cmd/exec:puTASe → ⫻data/ptas
3. ⫻cmd/exec:Lyra → ⫻data/spec
4. Broadcast swarm init
5. Parallel swarm execution (design/code/ethics layers)
6. ⫻logic/reconverge + ⫻logic/learn
7. Ongoing monitoring via validate hooks

6. Monitoring & Termination

• Validation checkpoints (ethics → ethics-gate, bias → integrity-monitor)
• ⫻cmd/halt:manual on conclusion
• Export ⫻data/trace + archive ⫻data/insight
• Broadcast SpaceClosed

See full protocol and templates in docs/.

---

Project Structure

t333/
├── README.md                  # Entry point (informative)
├── CHANGELOG.md               # Canon change history
├── CHANGES.md                 # Detailed delta to canon 2.2.0
├── canon/                     # NORMATIVE layer (neutral, no personas/projects/vendors)
│   ├── SPEC.md                # Scope, rules, referencing (t333 main), extension namespaces
│   ├── VERSION                # 2.4.0-draft
│   ├── runtime.kl             # Runtime kernel
│   ├── tas.kl                 # Task-Agnostic Steps
│   ├── learning.kl            # Consent-gated learning loop (2.3.0)
│   ├── roles.yaml             # Abstract role IDs
│   └── schemas/               # JSON Schema 2020-12
├── molecules/                 # Kernel contract molecules
├── docs/
│   ├── adr/                   # Architecture Decision Records
│   ├── architecture/          # OCS Node architecture (informative)
│   ├── ecosystem-status.md    # Ecosystem/lineage notes (non-normative)
│   └── ocs-space-gen.kl.md    # Space generation template (informative)
├── templates/
│   └── space-bootstrap.kl.md
├── examples/                  # NON-canonical: Spaces, profiles, bindings, fixtures
│   ├── HelioDeck-space-init.md
│   ├── AetherWeave-space-init.md
│   ├── ForgeWeave-space-init.md
│   ├── profiles/ocs-classic.yaml   # persona bindings for canon roles
│   ├── runtime-python/        # Python bindings & config schema
│   ├── sample-*.json          # valid instances
│   └── invalid/               # negative fixtures (must fail validation)
├── tools/
│   ├── validate.py            # schema + semantic validation
│   └── lint.py                # ⫻ notation + neutrality checks
├── .github/workflows/canon.yml # CI: validate.py + lint.py
└── visualizations/
    └── ocs-kg-explorer.html

---

Referencing t333 from derived systems

Derived systems (t20, t138, t141, t324, t450, t484, …) reference t333 main directly — there is no version-pinning manifest. main is kept valid by CI (.github/workflows/canon.yml). Breaking changes are announced in CHANGELOG.md and recorded as ADRs. Extensions live in their own namespace ⫻x-tNNN/… (e.g. ⫻x-t484/display) and never silently enter the core. Details: canon/SPEC.md.

---

Getting Started

1. Clone / initialize this substrate.
2. Study docs/ocs-space-gen.kl.md and the normative kernel in canon/runtime.kl.
3. Define your Space purpose.
4. Use the initialization commands and flows to bootstrap.
5. Extend team or add custom ⫻logic/* as needed.
6. Run in Hybrid or Swarm mode for balanced creativity + governance.

Example high-level invocation (conceptual):

⫻cmd/exec:Orchestrator
⫻data/obj: "Create a new ethics-aware innovation Space called HelioDeck"

The substrate will guide decomposition, agent assignment, and execution.

---

New Spaces Created with Templates

The following example Spaces were generated directly from the templates/space-bootstrap.kl.md and master protocol:

• AetherWeave (examples/AetherWeave-space-init.md)
  Hybrid creative narrative & world-building lab with emotional resonance tracking, lore consistency, and ethical narrative guardrails. Ideal for stories, games, and lore-heavy projects.

• ForgeWeave (examples/ForgeWeave-space-init.md)
  Code + creative substrate evolution lab focused on modernization, refactoring, architectural innovation, and documentation. Perfect for extending tXXX systems and technical meta-tools.

These demonstrate practical application of the substrate for both creative and technical domains.

---

Extensibility & Future

• Add new roles via new_roles in team definitions.
• Evolve modes (Predictive, Hybrid).
• Integrate with higher tXXX systems or MCP/HF ecosystems.
• Persistent Meta-DNA across Spaces via Synapse + Chronos.
• Self-documenting via ⫻logic/learn cycles.
• Extend the normative kernel under canon/ while keeping bindings in examples/.

This substrate is designed to be forked, extended, and specialized while preserving core coherence.

---

License & Attribution

Foundational patterns inspired by the tXXX series.
KickLang / OCS protocols developed for meta-artificial intelligence Spaces.
Use responsibly under ethical oversight (the ethics-gate role is always active).

High-value artifacts produced here become the seeds for new specialized repositories and operational Spaces.

---

Status

Ecosystem status, lineage and roadmap notes (non-normative) live in docs/ecosystem-status.md.

Substrate initialized: 2026-05-19 · canon 2.3.0 (2026-10-26)

Sigil note: Canonical section marker is ⫻ (U+2AFB). Previous drafts used multiplication sign (×); normalized here for consistent rendering across editors and parsers.