ariffazil/COPILOT_ZEN_ROUTER

🎯 COPILOT ZEN ROUTER — The Meta-Mesa

DITEMPA BUKAN DIBERI — One skill to route them all. The mesa above the terrain sees every path.

How Copilot Uses This

When you receive a task, check this taxonomy. Each layer lists exact skill names — load ONLY the matching skill(s). Never load more than 5 domain skills at once. The 9 BOOTSTRAP skills are pre-loaded.

---

L0 · BOOTSTRAP — Always Loaded (9 skills)

These fire at session start via BOOTSTRAP_MANIFEST.json. You already have them.

| # | Skill | Kernel Verb | What It Does | |---|-------|-------------|--------------| | 1 | kernel-bind | arif_init | Session governance binding | | 2 | observe-ground | arif_observe | Evidence before narrative | | 3 | route-dispatch | arif_route | Right organ for right intent | | 4 | memory-manage | arif_memory | Memory lifecycle discipline | | 5 | verify-gate | arif_verify | 4-gate verification | | 6 | audit-seal | arif_seal | Decision logging + sealing | | 7 | know-physics | arif_observe | Physical law veto | | 8 | know-math | arif_verify | Mathematical reasoning | | 9 | know-language | arif_compose | Linguistic competence |

---

L1 · GOVERNANCE — Constitutional Floor Enforcement

Load when: judging, sealing, floor-checking, authority questions, constitutional matters.

| Skill | Trigger | Use Case | |-------|---------|----------| | ASI-agent-invariants | "is this safe", "check floors", "constitutional" | 10 invariants + 12 governance rules | | ASI-constitutional-reasoning | "apply F1-F13", "floor check", "governed action" | F1-F13 reasoning before any governed action | | ASI-mcp-governor | "MCP routing", "tool authority", "governed MCP" | F1-F13 gated MCP routing | | APEX-humility-godel | "am I overconfident", "confidence check", "Gödel" | Self-critique before SEAL-grade claims | | APEX-fff-loop-protocol | "audit loop", "FFF loop", "recursive audit" | 5-pass recursive audit sequence | | APEX-quantum-eureka | "insight", "contradiction", "eureka moment" | Cross-domain synthesis | | delta-omega-psi-multimodal-cognition | "multimodal", "image + data", "cross-modal" | Δ·Ω·Ψ multimodal cognition rules |

When in doubt between these: Start with ASI-agent-invariants, escalate to ASI-constitutional-reasoning.

---

L2 · SENSE — Observation & Evidence

Load when: searching web, fetching URLs, grounding claims, observing state.

| Skill | Trigger | Use Case | |-------|---------|----------| | AGI-multimodal-bridge | "analyze image + data", "visual + textual evidence" | Multi-modal reasoning bridge | | FORGE-spatial-grounding | "where am I", "VPS location", "spatial context" | Geo-awareness for CLI operations | | geox-grounding | "earth evidence", "geo context", "basin" | Geological grounding layer | | FLAME-router | "route to FLAME", "free-tier model" | FLAME inference routing | | FLAME-operator | "FLAME health", "FLAME hit rate", "model diagnostics" | FLAME health monitoring |

Tools for SENSE (MCP — no skill needed):

• forge_fetch (mode=search) — web search via SearxNG
• forge_fetch (mode=readable) — URL content extraction
• brave_web_search — Brave Search API
• arif_observe — kernel sense verb

---

L3 · THINK — Reasoning & Planning

Load when: reasoning, planning, analyzing, summarizing, binding skills.

| Skill | Trigger | Use Case | |-------|---------|----------| | AGI-plan-dag | "multi-step plan", "DAG", "execution graph" | Build execution DAGs with checkpoints | | AGI-explorer-intelligence | "explore", "hypothesize", "falsify", "verify" | OBSERVE→HYPOTHESIZE→FALSIFY→VERIFY loop | | ASI-skill-binding | "discover skill", "bind skill", "compose skills" | Cross-organ skill discovery + composition | | ASI-summarize | "summarize", "condense", "TL;DR" | Summarize long content | | ASI-context-window-mgr | "context full", "compress", "token budget" | Context window lifecycle management | | AGI-claude-xml-structured-reasoning | "structured reasoning", "XML boundaries" | XML-tagged structured reasoning | | AGI-codex-chain-of-thought | "stepwise planning", "chain of thought" | Private stepwise CoT for Codex-style tasks | | causal555-pywhy | "causal inference", "counterfactual", "do-calculus" | Causal reasoning with PyWhy | | atlas333-cognitive-geometry | "cognitive geometry", "thought geometry" | Cognitive geometry analysis |

Tools for THINK (MCP — no skill needed):

• arif_think — kernel reasoning verb (modes: reason, reflect, verify, plan)
• sequential-thinking — MCP sequential thinking tool

---

L4 · ACT — Execution (FORGE Skills — 62 total)

L4a · GIT & GITHUB

Load when: git operations, PRs, commits, GitHub workflows.

| Skill | Trigger | |-------|---------| | FORGE-github-ops | "git push", "PR", "branch", "clone", "remote" | | FORGE-github-workflow | "CI/CD", "GitHub Actions", "workflow yaml" | | FORGE-pr-review | "review PR", "code review", "pull request review" | | FORGE-pr-governance | "governed PR", "PR governance", "approval gate" | | FORGE-precommit-review | "pre-commit", "before commit", "commit check" | | FORGE-precommit-gate | (loaded with precommit-review) | | code-review | "review this code", "code quality" |

L4b · INFRA & DOCKER

Load when: VPS operations, Docker, infrastructure, service management.

| Skill | Trigger | |-------|---------| | FORGE-vps-docker | "Docker", "container", "docker compose", "docker ps" | | FORGE-vps-runbook | "restart service", "systemctl", "VPS runbook" | | FORGE-infra-guardian | "Caddy", "SSL", "Cloudflare", "DNS", "tunnel" | | FORGE-infra-crons | "cron job", "crontab", "scheduled task" | | FORGE-docker-entropy | "docker cleanup", "orphan container", "docker prune" |

L4c · MCP & SKILLS

Load when: MCP server/tool work, skill creation/audit.

| Skill | Trigger | |-------|---------| | FORGE-mcp-ops | "MCP server", "MCP tool", "MCP transport" | | FORGE-mcp-lifeguard | "MCP health", "MCP down", "MCP recovery" | | FORGE-mcp-smoke-test | "test MCP", "MCP probe", "MCP smoke" | | FORGE-mcp-federation-ops | "federation MCP", "cross-organ MCP" | | FORGE-mcp-a2a-agentic | "A2A agent", "agent card", "agent mesh" | | FORGE-fastmcp | "FastMCP", "build MCP server", "Python MCP" | | FORGE-skill-creator | "create skill", "new skill", "forge skill" | | FORGE-skill-linter | "lint skill", "skill quality", "skill audit" |

L4d · WEB & SITES

Load when: building/deploying websites, frontend work.

| Skill | Trigger | |-------|---------| | FORGE-agentic-web-builder | "deploy site", "site down", "audit pages", "404" | | FORGE-nextjs-mastery | "Next.js", "React SSR", "Vercel" | | FORGE-react-spa-discipline | "React", "SPA", "Vite", "frontend app" | | FORGE-tailwind-tokens | "Tailwind", "CSS tokens", "design system" | | AGI-web-optimization | "SEO", "llms.txt", "web optimization" |

L4e · SECURITY

Load when: secrets, tokens, security scans, injection defense.

| Skill | Trigger | |-------|---------| | FORGE-secret-hygiene | "secret leak", "token check", "vault hygiene" | | FORGE-sct-federation-ingress | "SCT", "session token", "federation ingress" |

L4f · CI/CD & DEPLOY

Load when: CI pipelines, Docker builds, deployment.

| Skill | Trigger | |-------|---------| | FORGE-cicd-docker-deploy | "CI/CD pipeline", "Docker build", "deploy pipeline" | | FORGE-ci-diagnose | "CI broken", "workflow failed", "GitHub Actions fail" |

L4g · CODE ANALYSIS

Load when: analyzing repos, code quality, developer metrics.

| Skill | Trigger | |-------|---------| | FORGE-code-analysis | "analyze repo", "developer metrics", "code quality" | | FORGE-repo-intelligence | "repo overview", "repository map", "codebase" | | FORGE-context-compress | "compress output", "large log", "truncate" | | FORGE-context-compressor | (loaded with context-compress) |

L4h · OPS & MONITORING

Load when: monitoring, federation health, drift detection.

| Skill | Trigger | |-------|---------| | FORGE-route-least-power | "least power", "route optimization" | | FORGE-verify-runtime | "verify runtime", "runtime check", "SOT check" | | FORGE-federation-orchestrator | "federation orchestrate", "cross-organ" | | FORGE-federation-manifest | "federation manifest", "organ manifest" | | FORGE-telemetry-watchdog | "telemetry", "watchdog", "monitoring" | | FORGE-model-monitor | "model monitor", "LLM metrics" | | FORGE-symlink-audit | "symlink audit", "broken links" |

L4i · DOCS & TRUTH

Load when: documentation, README, cross-repo docs.

| Skill | Trigger | |-------|---------| | FORGE-readme-truth-check | "README check", "SOT stamp", "doc staleness" | | FORGE-cross-repo-doc-zen | "cross-repo docs", "doc consistency" | | FORGE-governance-jsonld | "JSON-LD", "governance context", "linked data" |

L4j · AGENT ORCHESTRATION

Load when: spawning sub-agents, cross-agent work.

| Skill | Trigger | |-------|---------| | FORGE-subagent-spawn | "spawn agent", "sub-agent", "parallel agent" | | FORGE-cross-agent-handoff | "handoff", "transfer task", "delegate" | | FORGE-onboarding | "onboarding", "new agent setup" | | FORGE-t3a-binding-matrix | "T3A matrix", "tool binding", "authority matrix" |

L4k · INCIDENT RESPONSE

Load when: incidents, outages, escalation.

| Skill | Trigger | |-------|---------| | FORGE-incident-escalation | "escalate", "incident", "SEV1" | | FORGE-incident-triage | "triage incident", "assess severity" | | FORGE-issue-triage | "triage issue", "GitHub issue", "prioritize" |

L4l · DATA & STORAGE

Load when: database, Redis, Qdrant, Postgres.

| Skill | Trigger | |-------|---------| | FORGE-postgres-schema-design | "Postgres schema", "database design", "migration" | | FORGE-redis-qdrant-integration | "Redis", "Qdrant", "vector store", "cache" | | FORGE-data-compression | "compress data", "archive", "storage optimization" |

L4m · SPECIALIZED EXECUTION

| Skill | Trigger | |-------|---------| | FORGE-did-web-identity | "DID", "decentralized identity", "WebID" | | FORGE-google-workspace | "Google Docs", "Gmail", "Google Drive" | | FORGE-kimi-code | "Kimi", "Moonshot AI", "Kimi agent" | | FORGE-grok-profile | "Grok profile", "X AI", "Grok config" | | FORGE-fastapi-api-builder | "FastAPI", "Python API", "REST endpoint" | | FORGE-vault999-witness | "VAULT999 witness", "seal chain witness" | | FORGE-seal-a-close | "seal session", "close session", "session handoff" | | FORGE-visual-qa-w3 | "visual QA", "screenshot audit", "W3 witness" | | forge-document-intelligence | "PDF", "OCR", "document scan", "extract text" | | FORGE-well-boundary-repair | "WELL boundary", "medical boundary" | | FORGE-telegram-audit | "Telegram audit", "Hermes Telegram" | | FORGE-mcp-gui | "MCP GUI", "MCP inspector", "MCP UI" |

---

L5 · SEAL — Memory & Vault

Load when: sealing, vault operations, session closure.

| Skill | Trigger | |-------|---------| | ASI-session-seal-copilot | "seal session", "end session", "handoff" (Copilot specific) | | ASI-session-seal | "seal session", "end session" (general) | | audit-seal | "audit seal", "vault seal", "seal receipt" | | AUDIT-recursive-audit | "audit skills", "skill overlap", "skill portfolio" | | AUDIT-drift-detector | "check drift", "manifest drift", "registry drift" |

---

L6 · META — Self-Improvement & Skill Management

Load when: managing skills, self-audit, improvement cycles.

| Skill | Trigger | |-------|---------| | AUDIT-skill-atlas | "skill atlas", "skill inventory", "gap detection" | | AGI-skill-unification | "skill unity", "skill mesh", "multi-harness skills" | | RSI-recursive-improvement | "RSI cycle", "self improve", "recursive improve" | | ASI-drift-watch | "drift watch", "SOT drift", "source drift" | | ASI-fabrication-prevention | "verify existence", "artifact check", "no fabrication" | | ASI-observability | "observability", "telemetry", "skill metrics" | | check-work | "check my work", "verify output" | | create-skill | "create skill", "bootstrap skill" |

---

L7 · DOMAIN — Organ-Specific Intelligence

GEOX (Earth Intelligence)

| Skill | Trigger | |-------|---------| | geox-grounding | "basin context", "geological evidence" | | geox-production-cockpit | "GEOX dashboard", "production cockpit" |

Tools (MCP — no skill needed): geox_basin, geox_petrophysics, geox_seismic_*, geox_well_*, geox_prospect, geox_claim, etc.

WEALTH (Capital Intelligence)

| Skill | Trigger | |-------|---------| | wealth-claim-state | "wealth claim", "capital claim", "financial state" |

Tools (MCP — no skill needed): capital_primitive, capital_health, capital_market, capital_wisdom, etc.

WELL (Human Readiness)

Tools (MCP — no skill needed): well_assess_homeostasis, well_validate_vitality, well_classify_substrate, etc.

---

L8 · KNOWLEDGE — Reference Substrate

Always available, never "loaded" — these are the knowledge floor.

| Skill | Purpose | |-------|---------| | knowledge | Federation knowledge substrate | | docs | Documentation reference | | runtime | Runtime environment reference | | scripts | Script reference | | substrate | Substrate reference | | reflective | Reflection reference |

---

L9 · SPECIALIZED — Niche Domains

Load only when the task explicitly matches these domains.

| Skill | Trigger | |-------|---------| | aaa-pdf-voice-protocol | "PDF voice", "voice protocol", "audio PDF" | | AGI-dream-engine | "dream engine", "memory consolidation", "nightly seal" | | AGI-emd-decode | "EMD decode", "decode upstream output" | | AGI-emd-encode | "EMD encode", "encode observation" | | AGI-emd-metabolize | "EMD metabolize", "memory promotion" | | AGI-entropy-lock-prime | "entropy lock", "ΔS gate", "entropy reduction" | | AGI-hermes-system-prompt-voice | "Hermes voice", "system prompt style" | | AGI-nusantara-substrate | "Nusantara", "cultural dignity", "sovereignty lens" | | ASI-agentic-architecture | "design agent", "agent architecture" | | ASI-agentic-governance | "AAA governance design", "AREP declaration" | | ASI-fabrication-prevention | "no hallucination", "verify before claim" | | apex-formal-constitution | "formal constitution", "constitutional law" | | apex_audit_coverage_check | "audit coverage", "test coverage audit" | | apex_authority_check | "authority check", "who can do this" | | apex_floor_check | "floor check", "F1-F13 check" | | apex_reversibility_test | "reversibility", "can we undo" | | apex_scope_check | "scope check", "blast radius" | | apex_tool_approval_gate | "tool approval", "register tool" | | apex_verdict_hold | "HOLD verdict", "pause action" | | apex_verdict_seal | "SEAL verdict", "approve action" | | asi_evidence_tier_express | "evidence tier", "evidence classification" | | asi_intent_hear | "intent hearing", "understand intent" | | asi_interface_adapt | "interface adaptation", "adapt output" | | asi_position_contrast | "position contrast", "counter-position" | | asi_tone_read | "tone reading", "emotional tone" | | asi_uncertainty_signal | "uncertainty", "confidence interval" | | causal555-pywhy | "causal inference", "do-calculus" | | federation-connect-headscale | "Headscale", "Tailscale", "VPN mesh" | | federation-release-attestation | "release attestation", "deploy attest" | | HERMES-opencode-protocol | "Hermes OpenCode", "Hermes protocol" | | KERNEL-trinity-33 | "Trinity 33", "final repo map", "3-axis" | | warga | "warga", "citizenship", "AAA membership" | | xauusd-trading | "XAUUSD", "gold trading", "forex" | | code-review | "code review", "review this" | | help | "help", "what can I do" | | imagine | "imagine", "creative", "generate idea" |

---

ROUTING ALGORITHM — How to Select

1. Is this a BOOTSTRAP function? → Already loaded (L0)
2. Is this about constitutional floors/authority? → L1 GOVERNANCE
3. Is this about observing/searching/gathering? → L2 SENSE
4. Is this about reasoning/planning? → L3 THINK
5. Is this about executing/building? → L4 ACT (sub-route by domain)
6. Is this about sealing/vault/memory? → L5 SEAL
7. Is this about managing skills/self-improvement? → L6 META
8. Is this about a specific organ (GEOX/WEALTH/WELL)? → L7 DOMAIN
9. Is this a niche/specialized domain? → L9 SPECIALIZED

ANTI-PATTERN: Load multiple FORGE skills for one task. Pick the ONE most specific. Example:

• "deploy site" → ONLY FORGE-agentic-web-builder (not infra-guardian + vps-docker + cicd)
• "review PR" → ONLY FORGE-pr-review (not github-ops + precommit + code-review)

PANIC RULE: Still unsure? Load AUDIT-skill-atlas — it's the master skill inventory.

---

Orthogonal Design Principles

1. No overlap: Each skill has one clear domain. If two skills trigger on the same phrase, merge them.
2. Layered loading: Bootstrap → Governance → Domain → Seal. Never skip layers on high-stakes work.
3. Harness-aware: Skills tagged harness: copilot-cli are native. Others may still work — test first.
4. Minimal loading: ≤5 domain skills per task. The agent is smart enough without overloading.
5. Tool-first: Many MCP tools (geox_, capital_, forge_*) need NO skill — just call them directly.

---

DITEMPA BUKAN DIBERI — Forged 2026-07-28 for Copilot CLI as the single meta-mesa routing layer. This skill IS the orthogonal taxonomy. Load it whenever the task spans multiple domains or the right skill is unclear.