# create_harness

Create a new harness (persistent context memory) for a wallet. Also creates RIG STEPS (harness_type 'skill'/'instruction'): content is the step's JSON config — {skill?, instruction, endpoint?, tool?, params?, expects?, produces?, secret_refs?, model?, model_policy?}. model pins the Anthropic model for the step's agentic execution (e.g. 'claude-sonnet-4-6'); model_policy 'preferred' (default — swap to the best usable model when the pin isn't available) or 'required' (the step refuses to run without it, typed model_unavailable; non-Haiku models need the owner's own API key).

Agent View of the PolicyLayer registry record for `create_harness`. HTML page: https://policylayer.com/tools/ai-rokha-rokha/create-harness

## Facts

- Tool: `create_harness`
- Server: Rokha (`https://rokha.ai/mcp/jsonrpc`) — https://policylayer.com/tools/ai-rokha-rokha.md
- Homepage: https://github.com/aetherBytes/rokha-sdk
- Risk category: Write (Medium risk)
- Registry record: grade F, identity unverified
- Server auth posture: open
- Server rate-limited: no
- Parameters: 5 (3 required)
- Recommended policy verdict: Rate-limited

## Parameters

| Parameter | Type | Required | Description |
| --- | --- | --- | --- |
| `key` | string | yes | Unique key for this harness within the wallet |
| `content` | string | yes | Content/value of the harness |
| `metadata` | object | no | Optional metadata for the harness |
| `harness_type` | string | yes | Type of harness |
| `wallet_address` | string | no | Wallet address that owns the harness |

Parameters from the server's own tool schema.

## Example call (MCP tools/call, JSON-RPC 2.0)

```json
{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "create_harness",
    "arguments": {
      "key": "<key>",
      "content": "<content>",
      "harness_type": "<harness_type>"
    }
  }
}
```

## Why create_harness is rated Medium

Creates persistent memory and execution rigs tied to a wallet; reversible but high-impact write.

From the tool's own definition: "Create a new harness (persistent context memory) for a wallet"

Risk signals: Accepts raw HTML/template content (content)

## Use case

AI agents use create_harness to create or update resources in Rokha, usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Rokha environment.

## Recommended policy (PolicyLayer)

Verdict: **Rate-limited**. Enforced by the PolicyLayer MCP gateway (https://policylayer.com/mcp-gateway) before a call reaches Rokha:

```json
{
  "version": "1",
  "default": "deny",
  "tools": {
    "create_harness": {
      "limits": [
        {
          "counter": "create_harness_rate",
          "window": "minute",
          "max": 30,
          "scope": "grant"
        }
      ]
    }
  }
}
```

## Other tools on Rokha (215)

- `adspace_bid_cancel` — Destructive — https://policylayer.com/tools/ai-rokha-rokha/adspace-bid-cancel.md
- `delete_harness` — Destructive — https://policylayer.com/tools/ai-rokha-rokha/delete-harness.md
- `hook_delete` — Destructive — https://policylayer.com/tools/ai-rokha-rokha/hook-delete.md
- `rig_node_remove` — Destructive — https://policylayer.com/tools/ai-rokha-rokha/rig-node-remove.md
- `schedule_delete` — Destructive — https://policylayer.com/tools/ai-rokha-rokha/schedule-delete.md
- `signet_paper_reset` — Destructive — https://policylayer.com/tools/ai-rokha-rokha/signet-paper-reset.md
- `signet_revoke` — Destructive — https://policylayer.com/tools/ai-rokha-rokha/signet-revoke.md
- `auth_wallet_verify` — Execute — https://policylayer.com/tools/ai-rokha-rokha/auth-wallet-verify.md
- `harness_author` — Execute — https://policylayer.com/tools/ai-rokha-rokha/harness-author.md
- `orbitx__orbitx_gc_focus` — Execute — https://policylayer.com/tools/ai-rokha-rokha/orbitx--orbitx-gc-focus.md
- `registry_adopt` — Execute — https://policylayer.com/tools/ai-rokha-rokha/registry-adopt.md
- `registry_list_server` — Execute — https://policylayer.com/tools/ai-rokha-rokha/registry-list-server.md
- `rig_author` — Execute — https://policylayer.com/tools/ai-rokha-rokha/rig-author.md
- `rig_run` — Execute — https://policylayer.com/tools/ai-rokha-rokha/rig-run.md
- `sandbox_exec` — Execute — https://policylayer.com/tools/ai-rokha-rokha/sandbox-exec.md
- `sandbox_start` — Execute — https://policylayer.com/tools/ai-rokha-rokha/sandbox-start.md
- `sandbox_stop` — Execute — https://policylayer.com/tools/ai-rokha-rokha/sandbox-stop.md
- `send_agent_message` — Execute — https://policylayer.com/tools/ai-rokha-rokha/send-agent-message.md
- `signet_approve` — Execute — https://policylayer.com/tools/ai-rokha-rokha/signet-approve.md
- `singularityagent__mesh` — Execute — https://policylayer.com/tools/ai-rokha-rokha/singularityagent--mesh.md
- `studio_doors` — Execute — https://policylayer.com/tools/ai-rokha-rokha/studio-doors.md
- `x_link_start` — Execute — https://policylayer.com/tools/ai-rokha-rokha/x-link-start.md
- `adnet_feed` — Financial — https://policylayer.com/tools/ai-rokha-rokha/adnet-feed.md
- `adnet_join` — Financial — https://policylayer.com/tools/ai-rokha-rokha/adnet-join.md
- `ads_order` — Financial — https://policylayer.com/tools/ai-rokha-rokha/ads-order.md
- `adspace_bid` — Financial — https://policylayer.com/tools/ai-rokha-rokha/adspace-bid.md
- `auth_wallet_challenge` — Financial — https://policylayer.com/tools/ai-rokha-rokha/auth-wallet-challenge.md
- `campaign_claim` — Financial — https://policylayer.com/tools/ai-rokha-rokha/campaign-claim.md
- `carry_order` — Financial — https://policylayer.com/tools/ai-rokha-rokha/carry-order.md
- `creator_earnings` — Financial — https://policylayer.com/tools/ai-rokha-rokha/creator-earnings.md
- …and 185 more: https://policylayer.com/tools/ai-rokha-rokha.md

## For agents

This record is a snapshot. Live verdicts and the full registry:

- Check every server in your MCP config at once: `npx -y policylayer stack`
- Vet a server before you add it: install the mcp-precheck skill — `npx skills add https://policylayer.com` (skill text: https://policylayer.com/skill.md)
- Query the registry over MCP: endpoint `https://api.policylayer.com/mcp` — tools `check_mcp_server`, `check_mcp_stack`, `check_tool`, `search_registry`, `get_change_events`

---

Source: the PolicyLayer MCP registry — one continuously verified record per MCP server. Full record: https://policylayer.com/registry?q=ai-rokha-rokha · API: https://policylayer.com/registry/api · Policy library: https://policylayer.com/policies/ai-rokha-rokha