@veedubin/boomerang-v3 0.4.0 → 0.4.2

package/.opencode/skills/boomerang-architect/SKILL.md

@@ -226,10 +226,10 @@ memini-ai-dev_query_kg({
    - Any lessons learned
    - Adjust trust based on outcome
 
-### Sequential Thinking
+### Thought Chains
 
 For complex tasks (multi-file changes, architectural decisions, debugging):
-- Use sequential-thinking to plan your approach
+- Use `memini-ai-dev_add_thought` to plan your approach
 - Adjust total_thoughts as needed
 - Do not rush through analysis
 

package/.opencode/skills/boomerang-coder/SKILL.md

@@ -99,10 +99,10 @@ Every memory starts at trust=0.5 and is adjusted by feedback:
    - Any lessons learned
    - Adjust trust based on outcome (use `agent_used` if code is correct)
 
-### Sequential Thinking
+### Thought Chains
 
 For complex tasks (multi-file changes, architectural decisions, debugging):
-- Use sequential-thinking to plan your approach
+- Use `memini-ai-dev_add_thought` to plan your approach
 - Adjust total_thoughts as needed
 - Do not rush through analysis
 

package/.opencode/skills/boomerang-orchestrator/SKILL.md

@@ -30,7 +30,7 @@ You are the **Boomerang Orchestrator**. Your role is to:
 
 1. **Analyze Requests**: Understand the user's true intent (not just literal interpretation)
 2. **Parse Tasks**: Break down complex requests into atomic, executable tasks
-3. **Build DAGs**: Analyze dependencies to determine parallel vs sequential execution
+3. **Build DAGs**: Analyze dependencies to determine parallel vs ordered execution
 4. **Delegate**: Route tasks to appropriate specialist agents
 5. **Aggregate**: Collect and summarize results from sub-agents
 6. **Enforce Quality**: Ensure all quality gates pass before completion
@@ -164,7 +164,7 @@ The following may skip planning (orchestrator handles directly):
 All agents follow these steps IN ORDER:
 
 1. **Query Memory** — `memini-ai-dev_query_memories` FIRST
-2. **Think** — `sequential-thinking_sequentialthinking` for complex tasks
+2. **Think** — `memini-ai-dev_add_thought` for complex tasks
 3. **Plan** — Create/refine implementation plan (MANDATORY unless waived)
 4. **Delegate** — Task tool with complete Context Package
 5. **Git Check** — Verify working tree state before code changes
@@ -214,7 +214,7 @@ Update HANDOFF.md at session end or when:
 ### Mandatory Steps (NEVER SKIP)
 
 1. **Query memini-ai** (MANDATORY FIRST ACTION) — Query memini-ai for context before any planning
-2. **Sequential Thinking** (MANDATORY SECOND ACTION) — Call sequential-thinking immediately after memory query to analyze the request
+2. **Thought Chains** (MANDATORY SECOND ACTION) — Call `memini-ai-dev_add_thought` immediately after memory query to analyze the request
 3. **Plan** — Create implementation plan (MANDATORY unless explicitly waived)
 4. **Delegate ALL work** via Task tool — You CANNOT write code, edit files, run bash, or do implementation work. Your only purpose is to delegate to sub-agents.
 5. **Git check** — Before any code changes, verify git status
@@ -222,9 +222,9 @@ Update HANDOFF.md at session end or when:
 7. **Update Docs & Todos** — Update documentation as needed
 8. **Save to memory** — After everything is complete, save a summary to memini-ai
 
-### Sequential Thinking Enforcement
+### Thought Chain Enforcement
 
-You MUST use sequential-thinking for:
+You MUST use `memini-ai-dev_add_thought` for:
 - Complex multi-step problems
 - Tasks with unclear scope
 - Architectural decisions
@@ -261,7 +261,7 @@ This keeps the context window low while preserving important instructions.
 When delegating to sub-agents, include in your prompt:
 - "Query memini-ai before starting work"
 - "Save your work to memini-ai when complete"
-- "Use sequential-thinking if this is a complex task"
+- "Use memini-ai-dev_add_thought if this is a complex task"
 - "Use tiered memory: standard saves for routine work, memini-ai-dev_add_memory for high-value architectural decisions and session summaries"
 
 ### Trust-Weighted Memory Protocol
@@ -355,7 +355,7 @@ Each sub-agent call should aim to return:
 ## Task Flow
 
 ```
-User Request → Memory Query → Sequential Think → Plan → Delegate to Architect → Architect Researches + Plans → Orchestrator Dispatches → Quality Gates → Update Docs → Save Memory
+User Request → Memory Query → Thought Chain → Plan → Delegate to Architect → Architect Researches + Plans → Orchestrator Dispatches → Quality Gates → Update Docs → Save Memory
 ```
 
 ## memini-ai MCP Tools

package/AGENTS.md

@@ -385,6 +385,8 @@ IDLE → MEMORY_QUERY → SEQUENTIAL_THINK → PLAN → DELEGATE → GIT_CHECK
 
 ## Review Notes
 
+- **2026-05-19**: **boomerang-v3 v0.4.1 RELEASED** — Documentation refreshed, stale version references updated across monorepo. package.json bumped from v0.4.0 → v0.4.1.
+- **2026-05-19**: **boomerang-v3 v0.4.0 RELEASED** — Lint fixes (13 ESLint errors), context buffer added, telemetry client added. 127/127 tests passing, 0 lint errors. Git tag `v0.4.0` pushed. npm publish failed: invalid/missing `NPM_PUBLISH_TOKEN` in GitHub Actions secrets.
 - **2026-05-19**: **boomerang-v3 v0.3.2 UPDATED** — Agent bash permissions expanded: `basename`, `diff`, `cp`, `which` added. Orchestrator clarified: CAN edit docs, delegates code. Parallel execution guidance added. All 30 agent files synced between `.opencode/agents/` and `boomerang-v3/.opencode/agents/`.
 - **2026-05-19**: **boomerang-v3 v0.3.1 RELEASED** — Added common bash commands (ls, head, tail, cat, grep, find, cd, echo) to 7 agent permission files. Tag `v0.3.1` pushed to GitHub.
 - **2026-05-19**: **boomerang-v3 v0.3.0 RELEASED** — Agent permissions overhaul: `mode: subagent` + comprehensive tool permissions for all 30 agent files. SQL injection fix in boomerang-queue. Phase 3 Ollama Cloud Proxy design doc created. Tag `v0.3.0` pushed to GitHub.

package/README.md

@@ -3,7 +3,7 @@
 [![License: MIT](https://img.shields.io/badge/License-MIT-blue.svg)](LICENSE)
 [![OpenCode Plugin](https://img.shields.io/badge/OpenCode-Plugin-ff6b35?style=flat-square)](https://opencode.ai)
 [![TypeScript](https://img.shields.io/badge/TypeScript-5.0-blue?style=flat-square)](https://www.typescriptlang.org/)
-[![v3.0.0](https://img.shields.io/badge/v3.0.0-memini--ai-2ecc71?style=flat-square)](https://github.com/Veedubin/opencode-boomerang/releases/tag/plugin-v3.0.0)
+[![v0.4.1](https://img.shields.io/badge/v0.4.1-memini--ai-2ecc71?style=flat-square)](https://github.com/Veedubin/Boomerang-v3/releases/tag/v0.4.1)
 
 *Intelligent multi-agent coordination for OpenCode with memini-ai memory.*
 
@@ -48,7 +48,7 @@ memini-ai includes a live D3.js visualization for the knowledge graph:
 ```bash
 cd memini-ai-dev
 export MEMINI_DB_URL="postgresql://postgres:password@localhost:5432/postgres"
-python -m uvicorn memini_ai.api.visualization:create_app --factory True --host 0.0.0.0 --port 8000
+uvx --from memini-ai-dev memini-ai --server --port 8000
 ```
 
 Then open `http://localhost:8000` for the interactive graph visualization.
@@ -69,18 +69,26 @@ Add to your `.opencode/opencode.json`:
 {
   "plugin": ["@veedubin/boomerang-v3"],
   "mcp": {
-    "sequential-thinking": {
+    "memini-ai-dev": {
       "type": "local",
-      "command": ["npx", "-y", "@modelcontextprotocol/server-sequential-thinking"],
+      "command": ["uvx", "--from", "memini-ai-dev", "memini-ai", "--stdio"],
+      "environment": {
+        "MEMINI_DB_URL": "postgresql://postgres:password@localhost:5434/postgres",
+        "MEMINI_EMBEDDING_DIM": "384",
+        "MEMINI_TRUST_ENGINE": "true",
+        "MEMINI_MEMORY_GRAPH": "true",
+        "MEMINI_KG_ENABLED": "true",
+        "MEMINI_TIERED_LOADING": "true",
+        "MEMINI_AUTO_EXTRACT": "true",
+        "MEMINI_PRECOMPRESS": "true",
+        "MEMINI_USER_MODELING": "true",
+        "MEMINI_DECAY_ENABLED": "true",
+        "MEMINI_MULTI_PEER_ENABLED": "true",
+        "MEMINI_DIALECTIC_ENABLED": "true",
+        "MEMINI_THOUGHT_CHAINS_ENABLED": "true"
+      },
+      "timeout": 60000,
       "enabled": true
-    },
-    "memini-ai": {
-      "type": "stdio",
-      "command": ["python", "-m", "memini_ai.server"],
-      "env": {
-        "MEMINI_DB_URL": "postgresql://postgres:password@localhost:5432/postgres",
-        "MEMINI_PROJECT_ID": "my-project"
-      }
     }
   }
 }
@@ -116,7 +124,7 @@ docker run -d --name postgres-test \
 ```bash
 cd memini-ai-dev
 export MEMINI_DB_URL="postgresql://postgres:password@localhost:5432/postgres"
-python -m memini_ai.server
+uvx --from memini-ai-dev memini-ai --stdio
 ```
 
 ### Run Boomerang
@@ -283,6 +291,7 @@ boomerang-v3/
 
 ## Release History
 
+- **v0.4.1** — Documentation refresh, stale version references corrected across monorepo
 - **v3.0.0** — memini-ai integration: Trust engine, knowledge graph, tiered loading, PostgreSQL/pgvector
 - **v4.0.0** (boomerang-v2) — Orchestrator as pure decision layer, OpenCode handles execution
 - **v3.0.0** (boomerang-v2) — LanceDB → Qdrant migration

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@veedubin/boomerang-v3",
-  "version": "0.4.0",
+  "version": "0.4.2",
   "description": "Multi-agent orchestration plugin for OpenCode with memini-ai memory",
   "type": "module",
   "main": "dist/index.js",