hivemind – Coding skill for AI agents

Hivemind: Claude Code as Orchestrator, opencode as Free Worker Swarm

Claude Code = brain (plans, reviews, merges). opencode = disposable workers on free models

(opencode/mimo-v2.5-free default; verified $0.00 per run).

Prerequisites (external dependency)

This skill is a thin orchestration layer over [opencode](https://opencode.ai), a

third-party CLI. It is not bundled — install and authenticate it yourself first:

| Requirement | Notes |

|---|---|

| Node.js >= 18 | The scripts use fetch and node:timers/promises. |

| opencode CLI on PATH | npm i -g opencode-ai (or the installer opencode documents). |

| An authenticated opencode account | opencode auth login. Workers run as your account. |

| Default model opencode/mimo-v2.5-free | A free tier offered by opencode, not by Anthropic. Availability, rate limits, and pricing are opencode's to change — override with --model at any time. |

| Windows only: OPENCODE_GIT_BASH_PATH | Point at C:\Program Files\Git\bin\bash.exe, set persistently. |

Data flow (read before installing)

Hivemind is opt-in and ships outside every domain bundle. **By default your code

leaves your machine:** the default model, opencode/mimo-v2.5-free, runs on

opencode's hosted service, so each worker sends its prompt, plus any files or diffs

you give it, to opencode's endpoints. Nothing on the worker side calls the Anthropic

API.

To keep code local, point --model at a model served on your own machine (for

example an Ollama or LM Studio provider configured in your opencode config). With a

hosted model, delegate only code you are allowed to send to that provider, and never

delegate secrets.

Setup

1. Put this skill folder wherever your agent loads skills from (e.g. ~/.claude/skills/hivemind).

2. Export HIVEMIND_HOME pointing at that folder — the bundled slash commands use it:

   export HIVEMIND_HOME="$HOME/.claude/skills/hivemind"
   

3. Copy the bundled assets into place:

• assets/commands/*.md -> ~/.claude/commands/ (the /hive, /oc, /swarm, ... entry points)
• assets/agents/*.md -> ~/.config/opencode/agent/ (the scout / coder / tester worker personas)

Both copies are optional: everything the commands do can be driven by invoking

scripts/oc-worker.mjs directly, and any opencode agent name works with --agent.

Runtime state (.runs/*.jsonl) is written inside this folder and is gitignored.

Components

| Path (relative to this skill dir) | Purpose |

|---|---|

| scripts/oc-worker.mjs | ONLY sanctioned way to invoke a worker. Hardened join point. |

| scripts/oc-status.mjs | Fleet progress from run logs (oc-status.mjs <run-id>) |

| scripts/oc-aggregate.mjs | Dedupe/synthesize N worker outputs; consensus findings first |

| scripts/bench/run-bench.mjs | Benchmark configs A (claude solo), B (opencode solo), C (orchestrated swarm) |

| scripts/bench/grader-prompt.md | Blind grading rubric (max 12 pts + PASS/FAIL gate) |

| scripts/oc-verify.mjs | Fan a worker's claims to a verifier on a DIFFERENT model; returns counts + needs_review |

| assets/commands/ | Slash-command entry points to copy into ~/.claude/commands/ |

| assets/agents/ | scout / coder / tester / verifier agent definitions for opencode |

Slash commands (ship in assets/commands/, copy to ~/.claude/commands/):

• /hive <task> - AUTO-ROUTER. Classifies task -> single worker, generic swarm, or template. Default entry point; prefer this over manual routing.
• /oc <task> - single worker delegation
• /swarm <task> - generic parallel swarm
• /review-panel <diff> - 4-lens parallel review (correctness/security/performance/style) + consensus aggregation
• /research-sweep <question> - 3-5 parallel research angles, synthesized
• /migration <task> - batched per-worktree migration workers + sequenced merge
• /test-fleet <target> - partitioned parallel test runs with safety checks

Worker agents (ship in assets/agents/, copy to ~/.config/opencode/agent/):

• scout - read-only research (no write/edit/bash)
• coder - implements one subtask in its worktree
• tester - runs tests only, never edits source

Invocation contract

node "<skill-dir>\scripts\oc-worker.mjs" [--agent scout|coder|tester] [--dir <path>] [--model <p/m>] [--timeout 900] [--run <id> --label <name>] [--json] "TASK TEXT"

Returns exactly ONE compact JSON line:

{ ok, result, tokens:{total,input,output,cache}, cost_usd, duration_ms, label, agent, model }

On failure: { ok:false, stage:"args"|"exec"|"api"|"parse"|"empty"|"schema", error } with stderr capped at 300 chars.

--json makes result a JSON value: the script extracts it from the reply, gives the model one

corrective retry, and fails with stage:"schema" if it still gets no parseable JSON.

--run <id> + --label <name> append lifecycle events (start/done/fail) to .runs/<id>.jsonl

inside this skill dir. Use them for EVERY swarm worker so progress is recoverable via

oc-status.mjs even after orchestrator context loss.

The script auto-manages the shared server: health-checks 127.0.0.1:4096, spawns opencode serve if dead, waits 5s, falls back to cold start. Workers are idempotent against their --dir; re-run once on ok:false before giving up.

HIVEMIND_SERVER_URL overrides that address (default http://127.0.0.1:4096). It must be a

valid URL with a numeric port; anything else fails fast with a single stage:"args" JSON line

rather than reaching the spawned process.

Verify instead of re-reading (do this before you trust a claim)

Re-checking every worker claim yourself costs more than the delegation saved. Send the

claims to a verifier instead:

node scripts/oc-verify.mjs --dir <same dir the worker saw> --claims findings.json --run <id>

Returns { confirmed, refuted, unsupported, unverified, needs_review[] }. Read

needs_review only — those are the refuted and unsupported claims. Confirmed claims are

counts, not prose, and do not come back into your context.

HARD RULES for this stage:

• The verifier MUST run on a different model from the worker. Same model = same blind

spots = a rubber stamp. The verifier defaults to opencode/nemotron-3.5-lightning-free

against the worker's default opencode/mimo-v2.5-free; pass --model if you change either.

• The verifier sees the FILES and the claims, never the worker's reasoning.
• It defaults to REFUTED when uncertain, because a false CONFIRMED is acted on unchecked.
• unverified > 0 means the verifier skipped claims. Treat those as unverified, not passed.
• Still spot-check a sample yourself. Two free models agreeing is evidence, not proof.

Free models are not always up

Free opencode/* models return provider 404s and empty bodies often enough that an

unattended swarm cannot depend on any one of them. Never hardcode a free model id: they

stop being free without warning. Pass --model explicitly for scheduled or bulk work, and

when a free model fails twice, follow the fallback ladder below rather than retrying.

Golden Rule (non-negotiable)

Raw opencode NDJSON streams must NEVER enter your context. All output arrives via the

script's single JSON line. Never pipe opencode run --format json directly into this

conversation; never re-implement what the script does.

Single worker flow (/oc)

For one read-only question or small delegation: run oc-worker.mjs without worktrees.

Read-only tasks may omit --agent/--dir. Summarize result for the user.

If files were written: show git diff before letting the user commit.

Swarm flow (multi-worker)

1. Decompose task into 2-5 INDEPENDENT subtasks (no shared files).

2. Writing workers get isolated worktrees FIRST: git worktree add ../<repo>-wt-N -b swarm/N.

3. Issue ALL worker invocations as PARALLEL Bash tool calls in ONE message.

4. Review every diff yourself (git diff main...swarm/N). YOU are the only merger.

5. Merge approved branches, remove worktrees, run tests.

6. Report table: subtask | agent | tokens | outcome + total worker tokens.

HARD RULES: workers never share directories; never delegate merging/reviewing;

escalate to your own Sonnet only when a free-model worker demonstrably fails twice.

Benchmarking

node scripts\bench\run-bench.mjs --repo <project> [--configs a,b,c] [--task 1-5]

Appends JSONL records (ts, config, tokens, cost, duration) to bench-results.jsonl.

Grade artifacts blind with grader-prompt.md (grader sees only task spec + output).

Configs: A=claude solo baseline, B=opencode solo, C=claude orchestrating 2 workers.

Fallback ladder (all flows)

1. Worker ok:false -> re-invoke once against the same dir.

2. Still failing -> orchestrator performs that subtask inline, marks it [orchestrator-sourced].

3. opencode entirely down (exec/api twice) -> announce, abandon workers, do the task directly.

Never let a swarm fail a task that Claude could have done itself.

Fleet patterns

Four reusable topologies ship as s

… (full content is in the download package)