hivemind-plugin

Use when detecting and guardrailing against premature completion claims — verifying that subagent "done" assertions are backed by fresh runtime evidence, enforcing dual-signal completion protocol (doer + verifier must agree), and catching hollow "it works" claims that skip verification. Triggers on: "verify completion", "completion detection", "dual-signal", "premature completion", "evidence before claims", "hollow done", "fake done", "completion guardrail", "verify before complete", "block completion until verified", "done but not verified", "did it actually work", "completion verification", "dual-signal completion", "two-agent completion", "fresh evidence requirement". NOT for loop mechanics (iterative-loop), delegation patterns (subagent-delegation-patterns), or quality gate triad orchestration (quality-gate-orchestration).

Use when making cross-cutting changes, cross-pane modifications, framework migrations, or breaking changes that span multiple layers or frameworks (GSD, BMAD, Hivemind, or other). Triggers on: "cross-cutting change", "cross-pane change", "multi-layer change", "framework migration", "breaking change across frameworks", "change impact analysis", "cross-framework change", "multi-pane impact", "test-first change ordering", "dependency ordering across layers", "consumer impact tracing", "interface-first change", "ordered change management". NOT for single-layer changes, single-framework refactors, or cosmetic edits. Framework-agnostic — works across GSD, BMAD, Hivemind, or any project governance framework.

Evaluates whether implementation evidence is sufficient to pass quality gates. Enforces an evidence hierarchy from live runtime proof (L1) down to documentation summaries (L5), and refuses gate passage when evidence is missing, mocked where integration is claimed, or insufficient for the gate type. Use during code review gates, phase audits, milestone verification, integration checks, and deployment readiness. Activates after gate-spec-compliance clears spec alignment — this is the terminal gate in the triad (lifecycle → spec → evidence). Triggers: "evidence check", "gate evidence", "verify runtime proof", "evidence truth", "is there proof this works", "evidence hierarchy", "gate truth", "runtime evidence", "integration evidence", "mock-only detection", "completion honesty", "gate passed", "gate failed". Terminal skill in the quality gate triad — if evidence PASSES, all 3 gates clear.

Internal quality gate that evaluates whether Hivemind harness implementations correctly participate in the runtime lifecycle — covering 9-surface mutation authority, CQRS boundaries, actor hierarchy, event-driven wiring, classification fit (src/ vs .opencode/ vs .hivemind/), and OpenCode SDK surface compliance. Synthesized from .planning/codebase/ARCHITECTURE.md (9-surface authority table) and ingested @opencode-ai/plugin SDK v1.14.44 from anomalyco/opencode (tool(), hook() signatures). Use when performing a lifecycle gate check, auditing harness module integration, verifying CQRS boundary compliance, checking delegation hierarchy constraints, evaluating tool/hook registration correctness, running a harness quality gate, validating plugin composition integrity, or running phase audit on src/ modules. Activates during code review of src/ files, phase audit, milestone verification, integration check, and deployment readiness workflows.

Spec compliance gate performing bidirectional traceability, gap detection (4 types), EARS acceptance criteria validation, and anti-pattern scanning. Use during code review gates, phase audits, milestone verification, and deployment readiness. Middle gate in the quality triad (lifecycle → spec → evidence). Routes to gate-evidence-truth on PASS; STOPS with gap report on FAIL. Includes remediation routing to hm-spec-driven-authoring, hm-test-driven-execution, and hm-debug for fix workflows. Triggers on: "spec compliance", "verify against spec", "gap analysis", "compliance gate", "phase audit gate", "acceptance criteria check", "spec-to-code", "deployment readiness", "triad gate", "spec gate middle", "quality triad".

Detects and fixes drift between AGENTS.md documentation and actual codebase state. Scans source files and .opencode/ directories, compares claims against reality, produces a structured drift report, then applies targeted edits. Triggers on: 'sync agents md', 'update AGENTS.md', 'fix agents md drift', 'AGENTS.md out of date', 'check agent instruction drift'. NOT for generic documentation writing or README refreshes.

Enforce pre-delegation authorization gates before agent dispatch. Use when setting up checkpoint gates, defining capability matrices, validating agent permissions, or approving a handoff boundary. NOT for orchestration execution, direct implementation, or generic task planning.

Routes requests about OpenCode meta-concepts (skills, agents, commands, tools) to specialist authors. Classifies intent, navigates step-by-step, reports back. Use when creating, auditing, stacking, or configuring any OpenCode meta-concept. Triggers on: 'create a skill', 'audit this skill', 'build an agent', 'set up a command', 'stack skills', 'configure OpenCode', 'fix skill trigger', '/hf-create', '/hf-audit', '/hf-stack'.

This skill should be used when the user asks to "create a skill", "audit this skill", "refactor skills", "doctor agent skills", "check skill quality", "fix frontmatter", "skill pattern selection", "TDD for skills", "cross-platform skill compatibility", or "score skill quality".

LOAD FIRST — session governance core. Use actual tools to discover session state, filter active/resumable sessions, query hierarchy, and aggregate cross-session data. No aspirational workflow — only real tool capabilities.

Use when running iterative refinement loops with bounded iteration control, entry/exit criteria enforcement, max-iteration guards, and checkpoint-based recovery between cycles. Triggers on: "iterative loop", "loop until complete", "iteration control", "loop with max iterations", "multi-round refinement", "progressive improvement", "retry-with-correction", "repeat until condition", "loop with exit criteria", "loop termination", "bounded iteration", "iteration guard", "checkpoint recovery in loop", "rerun with fixes", "fix → verify → repeat", "loop until criteria met", "iteration budget". NOT for wave-based parallel execution (wave-execution), guardrail against premature completion claims (completion-detection), or multi-agent coordination patterns (multi-agent-coordination).

Thực hiện nghiên cứu thị trường toàn diện với các công cụ phân tích đối thủ cạnh tranh, nghiên cứu xu hướng thị trường, nghiên cứu từ khóa SEO, khảo sát và insight khách hàng, và tổng hợp báo cáo thị trường. Dành cho marketing manager cần dữ liệu để ra quyết định. Dùng khi bạn nói "phân tích đối thủ", "nghiên cứu thị trường", "research từ khóa", "xu hướng ngành", "insight khách hàng", "báo cáo thị trường", "competitor analysis", "market research", "SEO research", "customer insight", "phân tích ngành".

Coordinate multiple AI agents through structured dispatch patterns. Use when dispatching subagents, running parallel agent tasks, organizing wave-based execution, designing multi-agent workflows, batching related agent delegations, avoiding over-delegation, setting checkpoint gates in agent loops, enforcing max-iteration limits, recovering orphan subagents, or choosing between parallel vs pipeline vs wave dispatch. NOT for single-agent delegation mechanics or quality-gate evaluation — those are separate skill domains.

Framework-agnostic guided workflow for configuring OpenCode agents, commands, and skills programmatically. Turn-based: Discover → Investigate → Collect → Proposal → Validate → Compile → Test → Save. Supports individual and batch configuration with cross-primitive conflict detection. Coexists safely with GSD, BMAD, Speckit, and other frameworks. Triggers on: 'configure agent', 'configure command', 'configure skill', 'batch configure', 'agent setup', 'set up agent', 'configure OpenCode primitives', 'update agent', 'modify agent', 'change agent', 'batch update', 'reconfigure', 'gatekeeping setup' .

Orchestrate quality gates as a fixed-sequence triad (lifecycle → spec compliance → evidence truth). Use during code review gates, phase audits, milestone verification, integration checks, and deployment readiness. Triggers on: "run quality gates", "gate orchestration", "quality triad", "full gate pass", "gate check before deploy", "pipeline gate", "gate sequence", "verify against gates", "phase audit gate", "triad verification", "quality gate pipeline", "route gates", "evidence hierarchy check", "gate decision matrix", "which gate to run first", "PASS FAIL routing", "gate remediation". Framework-agnostic — applies to any gate system with ordered quality stages. NOT for individual gate execution, coordination dispatch, or completion detection.

Discover, navigate, and aggregate session state across the project. Use when you need to: find active sessions, list all sessions, check session status, explore delegation hierarchy, trace parent-child chains, aggregate cross-session context, find related sessions by tool overlap, search session content, build a unified session view, resume an interrupted session, stack new work onto an existing session, or determine which sessions are resumable before dispatch. This is a read-side, CQRS-compliant skill — it discovers and reads session state without mutating it.

bun-pty pseudo-terminal integration: lazy-loading patterns, signal handling, and cross-platform PTY management for Bun runtime

@json-render/react by Vercel Labs — Generative UI framework that constrains AI output to structured JSON mapped to predefined React components. Used as the core rendering engine for the Hivemind GUI side-car dashboard.

Next.js 16.x stack reference — App Router architecture, server/client components, route handlers, proxy, cache components, Turbopack

Vitest testing framework API reference, patterns, and harness-specific testing conventions

Zod v4 TypeScript-first schema validation — complete API, patterns, and v3→v4 migration

Framework-agnostic subagent delegation patterns for OpenCode. Use when deciding which delegation tool to use: "task tool", "delegate-task", "execute-slash-command", "subagent dispatch", "parent session stacking", "delegation status check", "attach work to existing session", "child session delegation", "WaiterModel dispatch", "session stacking", "delegation monitoring", "check delegation status", "how to delegate to an agent", "subagent vs delegate-task", "async delegation", "stack onto session", "delegation anti-patterns", "permission-aware dispatch". Also triggers on: parallel tasks, complex async work, command dispatch with agent override, monitoring running delegations, or when the agent needs to choose between task tool and delegate-task for subagent work. NOT for simple one-step tool calls or single-file edits.

Use when designing or evaluating how AI frameworks extract and clarify user intent, when choosing between interactive probing vs passive context absorption, when user requests are vague and you need a systematic approach to surface requirements, or when comparing GSD discuss-phase, BMAD context absorption, or Hivemind iterative probing patterns. Triggers on: "how to extract user intent", "interactive probing vs passive absorption", "when to ask questions vs infer from context", "framework intent patterns", "clarify intent systematically", "intent clarity to probing depth", "too many questions anti-pattern", "question fatigue", "progressive intent extraction". NOT for teaching specific delegation tools (that's subagent-delegation-patterns), NOT for session discovery or resumption (that's session-foundation).

Use when executing parallel task groups with dependency ordering, decomposing work into execution waves, detecting file conflicts between parallel tasks, setting max-iteration limits on execution phases, or checkpointing multi-wave progress for recovery. Triggers: "wave execution", "parallel execution", "execute in waves", "wave-based dispatch", "dependency ordering", "wave dependency analysis", "checkpoint recovery", "max iteration guard", "file conflict detection", "parallel task ordering", "group tasks into waves", "conflict detection in parallel tasks", "premature completion", "over-parallelization". NOT for single-task dispatch, unstructured sequential execution, coordination patterns (multi-agent-coordination), or loop control (iterative-loop).

Routes requests about OpenCode meta-concepts (skills, agents, commands, tools) to specialist authors. Classifies intent, navigates step-by-step, reports back. Use when creating, auditing, stacking, or configuring any OpenCode meta-concept. Triggers on: 'create a skill', 'audit this skill', 'build an agent', 'set up a command', 'stack skills', 'configure OpenCode', 'fix skill trigger', '/hf-create', '/hf-audit', '/hf-stack'.

Formal naming convention for all Hivemind meta-concepts. Defines prefixes, lineage rules, and validation for hm-*, hf-*, gate-*, stack-* skills and agents. Use when creating any new skill or agent to ensure consistent, predictable, machine-verifiable naming. Triggers: naming convention, naming rules, prefix rules, how to name a skill, how to name an agent, naming validation, naming syndicate, check name compliance, validate skill name, validate agent name, name taxonomy, lineage prefix, what prefix should I use, correct prefix for, check naming convention. NOT for general project file naming or non-meta-concept naming.

This skill should be used when the user asks to "create a skill", "audit this skill", "refactor skills", "doctor agent skills", "check skill quality", "fix frontmatter", "skill pattern selection", "TDD for skills", "cross-platform skill compatibility", or "score skill quality".

Provides the complete oh-my-openagent architecture as a packed reference for agent plugin design. Use when referencing oh-my-openagent architecture, plugin system design, hook system patterns, circuit breaker design, session continuity patterns, or skill loader architecture. NOT for direct code implementation.

Complete OpenCode platform documentation covering agents, plugins, custom tools, SDK, permissions, skills, commands, configs, MCP servers, and models. Use when needing OpenCode-specific agents, SDK, plugin hooks, tool definitions, permission model, MCP server config, or platform architecture. NOT for general coding tasks or non-OpenCode framework lookup.

Complete tool capability matrix for the Hivemind ecosystem. Documents ALL tools available (native OpenCode + Hivemind custom + MCP provider), with per-tool descriptions, permission levels (allow/ask/ask), glob support, per-depth typical permissions, and per-lineage rules (hm STRICT, hf FLEXIBLE, gate INTERNAL, stack reference). Use when configuring agent permissions, designing delegation boundaries, auditing tool exposure, or deciding which tools a specialist agent needs. Triggers: tool permission, tool capability, tool matrix, what tools are available, agent tool configuration, delegation tool boundary, which tools can L2 agents use, lineage tool rules, hm lineage tools, hf lineage tools. NOT for general coding tasks or direct code implementation.

OpenCode SDK + Plugin internals for feature development, architecture, auditing, quality checking, test building, and gatekeeping. Contains BEYOND-DOCS expert knowledge extracted from source.

Vitest testing framework API reference, patterns, and harness-specific testing conventions

Internal quality gate that evaluates whether Hivemind harness implementations correctly participate in the runtime lifecycle — covering 9-surface mutation authority, CQRS boundaries, actor hierarchy, event-driven wiring, classification fit (src/ vs .opencode/ vs .hivemind/), and OpenCode SDK surface compliance. Synthesized from .planning/codebase/ARCHITECTURE.md (9-surface authority table) and ingested @opencode-ai/plugin SDK v1.14.28 docs (tool(), hook() signatures). Use when performing a lifecycle gate check, auditing harness module integration, verifying CQRS boundary compliance, checking delegation hierarchy constraints, evaluating tool/hook registration correctness, running a harness quality gate, validating plugin composition integrity, or running phase audit on src/ modules. Activates during code review of src/ files, phase audit, milestone verification, integration check, and deployment readiness workflows.

Execute GSD-style phase plans with wave-based parallelization and checkpoint recovery. Use when running a multi-plan phase, managing plan dependencies, recovering from mid-phase interruptions, when plans need to run in parallel, when execution needs checkpoint gates, or when phase work spans multiple sessions. Even when the user says "run the phase" or "execute the plan." Triggers: "execute phase", "run phase", "phase execution", "wave-based execution", "parallel plans", "checkpoint recovery", "plan dependencies". NOT for single-task execution or unstructured work.

Documents the complete `.hivemind/` state root structure — session continuity, delegation records, event journals, workflow config, planning persistence, and recovery paths. Use when navigating `.hivemind/` directories, reading/writing state files, recovering from interrupted sessions, understanding delegation records, or consuming Hivemind runtime state. NOT for mutating state directly — this is a read-only reference for hm-* agents.

Compress research findings into actionable artifacts with tiered reduction. Use when packing codebases, extracting interfaces, or producing validated reports. Stage 3 of the hm-research-chain pipeline. Consumes findings from hm-detective and evidence from hm-deep-research. Uses cached API signatures from hm-tech-stack-ingest for validation. NOT for raw data collection or unfiltered dumps.

Next.js 16.x stack reference — App Router architecture, server/client components, route handlers, proxy, cache components, Turbopack

Evaluates whether implementation evidence is sufficient to pass quality gates. Enforces an evidence hierarchy from live runtime proof (L1) down to documentation summaries (L5), and refuses gate passage when evidence is missing, mocked where integration is claimed, or insufficient for the gate type. Use during code review gates, phase audits, milestone verification, integration checks, and deployment readiness. Activates after gate-spec-compliance clears spec alignment — this is the terminal gate in the triad (lifecycle → spec → evidence). Triggers: "evidence check", "gate evidence", "verify runtime proof", "evidence truth", "is there proof this works", "evidence hierarchy", "gate truth", "runtime evidence", "integration evidence", "mock-only detection", "completion honesty", "gate passed", "gate failed". Terminal skill in the quality gate triad — if evidence PASSES, all 3 gates clear.

Spec compliance gate performing bidirectional traceability, gap detection (4 types), EARS acceptance criteria validation, and anti-pattern scanning. Use during code review gates, phase audits, milestone verification, and deployment readiness. Middle gate in the quality triad (lifecycle → spec → evidence). Routes to gate-evidence-truth on PASS; STOPS with gap report on FAIL. Includes remediation routing to hm-spec-driven-authoring, hm-test-driven-execution, and hm-debug for fix workflows. Triggers on: "spec compliance", "verify against spec", "gap analysis", "compliance gate", "phase audit gate", "acceptance criteria check", "spec-to-code", "deployment readiness", "triad gate", "spec gate middle", "quality triad".

Teaches agents how to compose specialist agents from user intent using XML markup grammar, step protocols, and structured return formats. Use when creating agents, designing agent definitions, composing multi-agent workflows, writing agent XML blocks, structuring execution flows, defining deviation rules, setting checkpoint protocols, or building structured return templates. NOT for direct code implementation.