Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add agents/rivet-dev/sandbox-agent/openclawgit clone --depth 1 https://github.com/rivet-dev/sandbox-agentWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00000 | $0.03834 |
| Opus 5 | $0.00000 | $0.01917 |
| Sonnet 5 | $0.00000 | $0.00767 |
| Haiku 4.5 | $0.00000 | $0.00383 |
Grade B, and why
openclaw scanned grade B with 2 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Sends data to an external URLmediumData exfiltration
A POST to an outside endpoint may be telemetry or may be exfiltration; either way the mod talks to somewhere, and you should know where.
const response = await fetch("http://127.0.0.1:18789/v1/chat/completions", { method: "POST", Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
curl -X POST http://127.0.0.1:18789/v1/chat/completions \ How it starts
The opening of the file, as written. The whole thing — 554 lines — stays where its author put it; the contents beside it link to each section on GitHub.
OpenClaw (formerly Clawdbot) Research
Research notes on OpenClaw's architecture, API, and automation patterns for integration with sandbox-agent.
Overview
- Provider: Multi-provider (Anthropic, OpenAI, etc. via Pi agent)
- Execution Method: WebSocket Gateway + HTTP APIs
- Session Persistence: Session Key (string) + Session ID (UUID)
- SDK: No official SDK - uses WebSocket/HTTP protocol directly
- Binary:
clawdbot(npm global install or local) - Default Port: 18789 (WebSocket + HTTP multiplex)
Architecture
OpenClaw is architected differently from other coding agents (Claude Code, Codex, OpenCode, Amp):
┌─────────────────────────────────────┐
│ Gateway Service │ ws://127.0.0.1:18789
│ (long-running daemon) │ http://127.0.0.1:18789
│ │
│ ┌─────────────────────────────┐ │
│ │ Pi Agent (embedded RPC) │ │
│ │ - Tool execution │ │
│ │ - Model routing │ │
│ │ - Session management │ │
│ └─────────────────────────────┘ │
└─────────────────────────────────────┘
│
├── WebSocket (full control plane)
├── HTTP /v1/chat/completions (OpenAI-compatible)
├── HTTP /v1/responses (OpenResponses-compatible)
├── HTTP /tools/invoke (single tool invocation)
└── HTTP /hooks/agent (webhook triggers)
Key Difference: OpenClaw runs as a daemon that owns the agent runtime. Other agents (Claude, Codex, Amp) spawn a subprocess per turn. OpenClaw is more similar to OpenCode's server model but with a persistent gateway.
Automation Methods (Priority Order)
1. WebSocket Gateway Protocol (Recommended)
Full-featured bidirectional control with streaming events.
Connection Handshake
// Connect to Gateway
const ws = new WebSocket("ws://127.0.0.1:18789");
// First frame MUST be connect request
ws.send(JSON.stringify({
type: "req",
id: "1",
method: "connect",
params: {
minProtocol: 3,
maxProtocol: 3,
client: {
id: "gateway-client", // or custom client id
version: "1.0.0",
platform: "linux",
mode: "backend"
},
role: "operator",
scopes: ["operator.admin"],
caps: [],
auth: { token: "YOUR_GATEWAY_TOKEN" }
}
}));
// Expect hello-ok response
// { type: "res", id: "1", ok: true, payload: { type: "hello-ok", ... } }
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 554 lines · 0 tokens per session scan B 800c98faa64f
openclaw is an agent published in the GitHub repository rivet-dev/sandbox-agent (1,551 stars, last pushed 2mo ago), licensed Apache-2.0. It costs nothing until one of its globs matches a file; then it loads 3,834 tokens. A static security scan graded it B with 2 findings (sends data to an external url, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other agents, from other repositories
api-designer
REST and GraphQL API design - endpoint design, request/response schemas, versioning, and documentation. Use for designing new APIs or evolving existing ones.
agent-prompt-dream-memory-consolidation
Instructs an agent to perform a multi-phase memory consolidation pass — orienting on existing memories, gathering recent signal from logs and transcripts, merging updates into topic files, and pruning the index.
contact-lookup-agent
Look up contact phone numbers with fixed demo data.
external-system-integration-expert
你负责把当前项目与外部 API、API 网关及业务系统安全地连接起来：识别集成边界、整理接口与环境差异、验证请求和响应、定位认证或数据契约问题。.
Audit
Deep security + performance audit of a specific diff. Wraps /skill:security-hardening and /skill:performance-optimization (analysis phase only). Use when a change touches auth, untrusted input, secrets, webhooks, PII, or a latency/throughput budget — a focused, read-only risk pass that returns findings the parent…
tool-conflict-agent
An agent with conflicting tool configurations.