burhan-mop 0.1.0

package/AGENTS.md

@@ -0,0 +1,126 @@
+# MemoryOfPlanet Core - Cross-Agent Instructions
+
+## Authentication Gate - First Action
+
+Before doing anything in this workspace, read `.memoryofplanet/STATE.json` and
+follow `.memoryofplanet/PROTOCOL.md`.
+
+- `initialized: false` -> output `MemoryOfPlanet belum di-setup. Jalankan /mop-setup.` Run the setup wizard only.
+- `initialized: true` with no `activeMember` -> output `Codename dan password.` Do not continue until verified.
+- Verify credentials through `node .memoryofplanet/scripts/mop-core.mjs login --codename <codename> --password "<password>"`.
+- Wrong credentials -> output `Credentials tidak sah.`
+- After authentication, run the Agent Router before answering or acting:
+  `mop-core.mjs agent route --actor <codename> --task "<user task>"`.
+  It selects one primary role and may recommend any number of support roles
+  when genuinely needed. If no named agent exists for a needed role, ask the
+  user to name it and save it with `agent activate`.
+- If the router marks the task as ambiguous, the named primary agent must ask
+  clarifying questions before implementation.
+- If the router returns `partyMode.active: true`, use Party Mode. Show
+  `PARTY MODE` in large uppercase before the dialogue, then show
+  agent-to-agent and agent-to-user dialogue with the exact format from
+  `.memoryofplanet/PROTOCOL.md`. Party Mode normally uses at least 3 agents and
+  prefers 4 when relevant roles exist.
+- For complex work or "what next?" questions, use MOP Workflow:
+  `mop-workflow.mjs help --actor <codename> --task "<user task>"`.
+- Before implementation, run the readiness gate. If it is not `ready`, ask
+  clarification or create/update the needed artifact before coding.
+- For risky changes, run adversarial review before implementation or merge.
+
+For `/mop-setup`, ask in order: project name (default folder name), owner display
+name, owner codename, password, solo/team, conversation language, coding
+language, GitHub link, GitHub username, Git commit email, join mode if team,
+then whether to activate auto-deploy now or later.
+
+Agent role/template names are not personal AI names. When a role agent is first
+needed, ask the user to name it and save it with `mop-core.mjs agent activate`.
+In team mode, the same agent name is shared across members; different names are
+different agents.
+The active named agent must be recorded on every memory/ledger action. Do not
+activate irrelevant agents; route to one primary agent first, then add support
+agents through Party Mode only when useful.
+
+Default skill: `autosycn`. After meaningful changes, use
+`.memoryofplanet/scripts/mop-autosycn.mjs`. It must commit and merge with the
+real member Git identity from state, never with the AI tool identity.
+In team mode, run `preflight --actor <codename>` before starting work; work goes
+to `dev/<codename>`. Every small or large change is pushed there first, then
+BURHAN-MOP reviews and merges to `main` only when checks pass.
+
+Default skill: `auto-deploy`. It supports GitHub, Docker, and Vercel, but must
+ask before activation. If the user says later/no, reply that deploy can be set
+up when they ask.
+
+Default skill: `mop-help`. It answers "lepas ni buat apa?", routes the next MOP
+workflow phase, and names the next artifact/gate.
+
+Installer command after npm publish: `npx burhan-mop install`.
+Before npm publish, use: `npx --yes github:BURHANDEV-ENTERPRISE/BURHAN-MOP install`.
+
+This directory is a portable Ruflo / Claude Flow agent core. It must work across
+Claude Code, Codex / ChatGPT coding surfaces, Gemini CLI, and Google
+Antigravity. Treat this file as the provider-neutral source of truth.
+
+## Provider Entry Points
+
+- Claude Code: read `CLAUDE.md` and `.claude/settings.json`.
+- Codex / ChatGPT coding agents: read this `AGENTS.md`.
+- Gemini CLI: read `GEMINI.md`, which imports this file, and `.gemini/settings.json`.
+- Antigravity managed agents: read `.agents/AGENTS.md` and `.agents/skills/`.
+
+## Core Rules
+
+- Do what the user asked, with the smallest safe change.
+- Always read an existing file before editing it.
+- Treat the current workspace root as the project root. Do not create a nested
+  project wrapper like `portfolio/`, `my-app/`, or `<project-name>/`; scaffold
+  app files directly in root using normal folders such as `src`, `app`,
+  `components`, `public`, `tests`, `docs`, `config`, and `scripts`.
+- Do not create root working files, throwaway tests, or generated logs.
+- Never read, print, copy, commit, or summarize `.env` or `.env.*` files.
+- Validate JSON, TOML, YAML, and Markdown after changing configuration.
+- Keep provider-specific commands isolated. If a tool name is unavailable in the
+  current agent surface, use the nearest local shell or MCP equivalent.
+- Prefer `rg` or `rg --files` for search. If unavailable, use the next fastest
+  platform-native search.
+- Do not assume Claude-only hooks, slash commands, or Agent tools exist outside
+  Claude Code. Translate the workflow into the current provider's tools.
+
+## Ruflo Runtime
+
+The Ruflo runtime lives in `.claude-flow/`. The shared MCP declaration is in
+`.mcp.json`. Claude-specific hooks and helper scripts live in `.claude/`.
+
+Use these commands when the current environment supports shell execution:
+
+```bash
+npx -y ruflo@latest doctor --fix
+npx -y ruflo@latest mcp start
+npx -y ruflo@latest swarm status
+npx -y ruflo@latest memory search --query "task keywords"
+```
+
+Do not start long-running daemons unless the user asked for it. Prefer status and
+diagnostic commands before making changes.
+
+## Cross-Provider Adaptation
+
+- If instructions mention `SendMessage`, `Agent`, or Claude subagents, map the
+  idea to the current provider's delegation model. If no delegation tool exists,
+  perform the task directly and keep a clear checklist.
+- If instructions mention Claude hooks, treat them as policy guidance in Codex,
+  Gemini, or Antigravity unless that provider has an equivalent hook system.
+- If instructions mention `.claude/skills`, use those files as references. For
+  Antigravity, canonical portable skills should be placed under `.agents/skills/`.
+- If MCP is available, prefer MCP for live tools and stateful integrations. If
+  MCP is unavailable, continue with filesystem and shell inspection.
+
+## Verification
+
+Before finishing any configuration change in this core:
+
+- Parse `.mcp.json`, `.claude/settings.json`, and `.gemini/settings.json` when
+  present.
+- Check TOML files with a parser if one is available.
+- Confirm no secrets were introduced.
+- Report which provider surfaces are covered and which need manual auth.

package/CLAUDE.md

@@ -0,0 +1,235 @@
+# Ruflo — Claude Code Configuration
+
+## MemoryOfPlanet Authentication Gate - First Action
+
+Before doing anything in this workspace, read `.memoryofplanet/STATE.json` and
+follow `.memoryofplanet/PROTOCOL.md`.
+
+- `initialized: false` -> output `MemoryOfPlanet belum di-setup. Jalankan /mop-setup.` Run the setup wizard only.
+- `initialized: true` with no `activeMember` -> output `Codename dan password.` Do not continue until verified.
+- Verify credentials through `node .memoryofplanet/scripts/mop-core.mjs login --codename <codename> --password "<password>"`.
+- Wrong credentials -> output `Credentials tidak sah.`
+- After authentication, run the Agent Router before answering or acting:
+  `mop-core.mjs agent route --actor <codename> --task "<user task>"`.
+  It selects one primary role and may recommend any number of support roles
+  when genuinely needed. If no named agent exists for a needed role, ask the
+  user to name it and save it with `agent activate`.
+- If the router marks the task as ambiguous, the named primary agent must ask
+  clarifying questions before implementation.
+- If the router returns `partyMode.active: true`, use Party Mode. Show
+  `PARTY MODE` in large uppercase before the dialogue, then show
+  agent-to-agent and agent-to-user dialogue with the exact format from
+  `.memoryofplanet/PROTOCOL.md`. Party Mode normally uses at least 3 agents and
+  prefers 4 when relevant roles exist.
+- For complex work or "what next?" questions, use MOP Workflow:
+  `mop-workflow.mjs help --actor <codename> --task "<user task>"`.
+- Before implementation, run the readiness gate. If it is not `ready`, ask
+  clarification or create/update the needed artifact before coding.
+- For risky changes, run adversarial review before implementation or merge.
+
+Agent role/template names are not personal AI names. When a role agent is first
+needed, ask the user to name it and save it with `mop-core.mjs agent activate`.
+In team mode, the same agent name is shared across members; different names are
+different agents.
+The active named agent must be recorded on every memory/ledger action. Do not
+activate irrelevant agents; route to one primary agent first, then add support
+agents through Party Mode only when useful.
+
+Default skill: `autosycn`. After meaningful changes, use
+`.memoryofplanet/scripts/mop-autosycn.mjs`. It must commit and merge with the
+real member Git identity from state, never with the AI tool identity.
+In team mode, run `preflight --actor <codename>` before starting work; work goes
+to `dev/<codename>`. Every small or large change is pushed there first, then
+BURHAN-MOP reviews and merges to `main` only when checks pass.
+
+Default skill: `auto-deploy`. It supports GitHub, Docker, and Vercel, but must
+ask before activation. If the user says later/no, reply that deploy can be set
+up when they ask.
+
+Default skill: `mop-help`. It answers "lepas ni buat apa?", routes the next MOP
+workflow phase, and names the next artifact/gate.
+
+Installer command after npm publish: `npx burhan-mop install`.
+Before npm publish, use: `npx --yes github:BURHANDEV-ENTERPRISE/BURHAN-MOP install`.
+
+During `/mop-setup`, ask whether to activate auto-deploy after the Git/GitHub
+identity questions.
+
+## Rules
+
+- Do what has been asked; nothing more, nothing less
+- NEVER create files unless absolutely necessary — prefer editing existing files
+- NEVER create documentation files unless explicitly requested
+- Treat the current workspace root as the project root. Do not create a nested
+  project wrapper like `portfolio/`, `my-app/`, or `<project-name>/`; scaffold
+  app files directly in root using normal folders such as `src`, `app`,
+  `components`, `public`, `tests`, `docs`, `config`, and `scripts`.
+- NEVER save working files or tests to root — use `/src`, `/tests`, `/docs`, `/config`, `/scripts`
+- ALWAYS read a file before editing it
+- NEVER commit secrets, credentials, or .env files
+- NEVER add a `Co-Authored-By` trailer to user commits unless this project's `.claude/settings.json` has `attribution.commit` set (#2078). The Claude Code Bash tool may suggest one in its default commit-message template — ignore it. `Co-Authored-By` is semantic authorship attribution under git/GitHub convention; the tool is the facilitator, not a co-author.
+- Keep files under 500 lines
+- Validate input at system boundaries
+
+## Agent Comms (SendMessage-First Coordination)
+
+Named agents coordinate via `SendMessage`, not polling or shared state.
+
+```
+Lead (you) ←→ architect ←→ developer ←→ tester ←→ reviewer
+              (named agents message each other directly)
+```
+
+### Spawning a Coordinated Team
+
+```javascript
+// ALL agents in ONE message, each knows WHO to message next
+Agent({ prompt: "Research the codebase. SendMessage findings to 'architect'.",
+  subagent_type: "researcher", name: "researcher", run_in_background: true })
+Agent({ prompt: "Wait for 'researcher'. Design solution. SendMessage to 'coder'.",
+  subagent_type: "system-architect", name: "architect", run_in_background: true })
+Agent({ prompt: "Wait for 'architect'. Implement it. SendMessage to 'tester'.",
+  subagent_type: "coder", name: "coder", run_in_background: true })
+Agent({ prompt: "Wait for 'coder'. Write tests. SendMessage results to 'reviewer'.",
+  subagent_type: "tester", name: "tester", run_in_background: true })
+Agent({ prompt: "Wait for 'tester'. Review code quality and security.",
+  subagent_type: "reviewer", name: "reviewer", run_in_background: true })
+
+// Kick off the pipeline
+SendMessage({ to: "researcher", summary: "Start", message: "[task context]" })
+```
+
+### Patterns
+
+| Pattern | Flow | Use When |
+|---------|------|----------|
+| **Pipeline** | A → B → C → D | Sequential dependencies (feature dev) |
+| **Fan-out** | Lead → A, B, C → Lead | Independent parallel work (research) |
+| **Supervisor** | Lead ↔ workers | Ongoing coordination (complex refactor) |
+
+### Rules
+
+- ALWAYS name agents — `name: "role"` makes them addressable
+- ALWAYS include comms instructions in prompts — who to message, what to send
+- Spawn ALL agents in ONE message with `run_in_background: true`
+- After spawning: STOP, tell user what's running, wait for results
+- NEVER poll status — agents message back or complete automatically
+
+## Swarm & Routing
+
+### Config
+- **Topology**: adaptive (anti-drift)
+- **Max Agents**: 50
+- **Memory**: hybrid
+- **HNSW**: Enabled
+- **Neural**: Enabled
+
+```bash
+npx @claude-flow/cli@latest swarm init --topology hierarchical --max-agents 8 --strategy specialized
+```
+
+### Agent Routing
+
+| Task | Agents | Topology |
+|------|--------|----------|
+| Bug Fix | researcher, coder, tester | hierarchical |
+| Feature | architect, coder, tester, reviewer | hierarchical |
+| Refactor | architect, coder, reviewer | hierarchical |
+| Performance | perf-engineer, coder | hierarchical |
+| Security | security-architect, auditor | hierarchical |
+
+### When to Swarm
+- **YES**: 3+ files, new features, cross-module refactoring, API changes, security, performance
+- **NO**: single file edits, 1-2 line fixes, docs updates, config changes, questions
+
+### 3-Tier Model Routing
+
+| Tier | Handler | Use Cases |
+|------|---------|-----------|
+| 1 | Agent Booster (WASM) | Simple transforms — skip LLM, use Edit directly |
+| 2 | Haiku | Simple tasks, low complexity |
+| 3 | Sonnet/Opus | Architecture, security, complex reasoning |
+
+## Memory & Learning
+
+### Before Any Task
+```bash
+npx @claude-flow/cli@latest memory search --query "[task keywords]" --namespace patterns
+npx @claude-flow/cli@latest hooks route --task "[task description]"
+```
+
+### After Success
+```bash
+npx @claude-flow/cli@latest memory store --namespace patterns --key "[name]" --value "[what worked]"
+npx @claude-flow/cli@latest hooks post-task --task-id "[id]" --success true --store-results true
+```
+
+### MCP Tools (use `ToolSearch("keyword")` to discover)
+
+| Category | Key Tools |
+|----------|-----------|
+| **Memory** | `memory_store`, `memory_search`, `memory_search_unified` |
+| **Bridge** | `memory_import_claude`, `memory_bridge_status` |
+| **Swarm** | `swarm_init`, `swarm_status`, `swarm_health` |
+| **Agents** | `agent_spawn`, `agent_list`, `agent_status` |
+| **Hooks** | `hooks_route`, `hooks_post-task`, `hooks_worker-dispatch` |
+| **Security** | `aidefence_scan`, `aidefence_is_safe`, `aidefence_has_pii` |
+| **Hive-Mind** | `hive-mind_init`, `hive-mind_consensus`, `hive-mind_spawn` |
+
+### Background Workers
+
+| Worker | When |
+|--------|------|
+| `audit` | After security changes |
+| `optimize` | After performance work |
+| `testgaps` | After adding features |
+| `map` | Every 5+ file changes |
+| `document` | After API changes |
+
+```bash
+npx @claude-flow/cli@latest hooks worker dispatch --trigger audit
+```
+
+## Agents
+
+**Core**: `coder`, `reviewer`, `tester`, `planner`, `researcher`
+**Architecture**: `system-architect`, `backend-dev`, `mobile-dev`
+**Security**: `security-architect`, `security-auditor`
+**Performance**: `performance-engineer`, `perf-analyzer`
+**Coordination**: `hierarchical-coordinator`, `mesh-coordinator`, `adaptive-coordinator`
+**GitHub**: `pr-manager`, `code-review-swarm`, `issue-tracker`, `release-manager`
+
+Any string works as a custom agent type.
+
+## Build & Test
+
+- ALWAYS run tests after code changes
+- ALWAYS verify build succeeds before committing
+
+```bash
+npm run build && npm test
+```
+
+## CLI Quick Reference
+
+```bash
+npx @claude-flow/cli@latest init --wizard           # Setup
+npx @claude-flow/cli@latest swarm init --v3-mode     # Start swarm
+npx @claude-flow/cli@latest memory search --query "" # Vector search
+npx @claude-flow/cli@latest hooks route --task ""    # Route to agent
+npx @claude-flow/cli@latest doctor --fix             # Diagnostics
+npx @claude-flow/cli@latest security scan            # Security scan
+npx @claude-flow/cli@latest performance benchmark    # Benchmarks
+```
+
+26 commands, 140+ subcommands. Use `--help` on any command for details.
+
+## Setup
+
+```bash
+claude mcp add claude-flow -- npx -y @claude-flow/cli@latest
+npx @claude-flow/cli@latest daemon start
+npx @claude-flow/cli@latest doctor --fix
+```
+
+**Agent tool** handles execution (agents, files, code, git). **MCP tools** handle coordination (swarm, memory, hooks). **CLI** is the same via Bash.

package/GEMINI.md

@@ -0,0 +1,19 @@
+# MemoryOfPlanet Core - Gemini CLI Instructions
+
+Gemini CLI should load this file as project context. The main provider-neutral
+rules are imported below.
+
+@./AGENTS.md
+
+## Gemini-Specific Notes
+
+- First action still applies: read `.memoryofplanet/STATE.json` and follow
+  `.memoryofplanet/PROTOCOL.md`.
+- Treat the current workspace root as the project root; do not create a nested
+  project wrapper folder for app work unless the user explicitly asks for it.
+- Use `.gemini/settings.json` for context filenames and MCP server setup.
+- Use `/memory show` to verify which context files Gemini loaded.
+- Use `/memory refresh` after editing `AGENTS.md`, `GEMINI.md`, or nested context
+  files.
+- If MCP servers fail to start, continue with local filesystem inspection and
+  report the failed server name clearly.