$ ~/registry/skill/openclaw-extensions-lobster

SKILL

lobster

Name: lobster
Author: openclaw

Write, understand, and debug code in the Lobster programming language.

Stars

★ 390,422

Source

GitHub

Updated

2026-09-25

// securityLow risk

Prompt only, no code
Open source
Community-trusted· 377.5k

Summary

Based on the provided materials, Lobster is classified as prompt-only and does not declare any required secrets, remote endpoints, or local system privileges. It mainly orchestrates multi-step workflows with approval checkpoints, so no high-risk behavior is evident by itself; overall risk is low, though effective power still depends on downstream tools it invokes.

CredentialsLow risk

The materials explicitly state that no keys or environment variables are required, and the README does not ask for tokens, API keys, or account credentials. For this skill itself, there is no evident credential collection, storage, or abuse surface.

Network egressLow risk

Both the metadata and checks indicate no remote endpoint host, and the README does not declare any fixed outbound destination. The skill itself does not appear to directly send user data to third-party network endpoints. If a pipeline later invokes other networked tools, that egress risk would come from those downstream tools rather than Lobster itself.

Code executionLow risk

The available materials classify it as prompt-only and do not declare any ability to spawn local processes, run scripts, or access system-level capabilities. The README describes workflow orchestration plus approval/pause-resume behavior, not a local code-execution interface.

Data accessLow risk

The materials do not declare direct read/write access to local files, databases, or other resources, nor do they request permissions beyond workflow orchestration. Examples such as email search and triage clearly depend on external tools being orchestrated, not on Lobster's own built-in data access.

Supply chainLow risk

The source is an open-source GitHub repository with auditable code and extremely high community adoption (about 377k stars), which strongly lowers supply-chain risk. The license and maintenance status are not clearly stated in the materials, but without concrete red flags this is not enough to elevate it to high risk.

Recommendations

Treat Lobster as an orchestration layer; the main security review should focus on the downstream tools it may invoke, including their permissions, endpoints, and data flows.
For pipelines with side effects such as send, delete, or sync, keep human approval gates mandatory and log approval events.
Before installation, verify consistency between the GitHub repository, published package, and skill name, and confirm the license and recent maintenance status.

Audited by: gpt-5.4 · 2026-06-16

// install

Copy the install command and let the AI configure it · recommended for beginners

Please install the "lobster" skill from askskill:
1. Download https://raw.githubusercontent.com/openclaw/openclaw/main/extensions/lobster/SKILL.md
2. Save it as ~/.claude/skills/lobster/SKILL.md
3. Reload skills and tell me it's ready

// download

Download SKILL.mdInstall manifest ↗

// examples

Generate a Lobster sample program

Prompt

Write a command-line todo app in Lobster that supports adding, listing, and deleting tasks, and add comments explaining the key syntax.

Expected

A runnable Lobster code sample with explanations of the syntax.

Explain a code error

Prompt

The following Lobster code throws an error. Find the issue, explain why it happens, and provide a corrected version:
[paste code here]

Expected

A clear explanation of the error and a corrected version of the code.

Port from another language

Prompt

Rewrite this Python code in Lobster and explain the main differences in variables, loops, and function syntax between the two languages:
[paste code here]

Expected

An equivalent Lobster implementation and notes on the syntax differences to watch for.

// docs

Lobster

Lobster executes multi-step workflows with approval checkpoints. Use it when:

User wants a repeatable automation (triage, monitor, sync)
Actions need human approval before executing (send, post, delete)
Multiple tool calls should run as one deterministic operation

When to use Lobster

User intent	Use Lobster?
"Triage my email"	Yes — multi-step, may send replies
"Send a message"	No — single action, use message tool directly
"Check my email every morning and ask before replying"	Yes — scheduled workflow with approval
"What's the weather?"	No — simple query
"Monitor this PR and notify me of changes"	Yes — stateful, recurring

Basic usage

Run a pipeline

{
  "action": "run",
  "pipeline": "gog.gmail.search --query 'newer_than:1d' --max 20 | email.triage"
}

Returns structured result:

{
  "protocolVersion": 1,
  "ok": true,
  "status": "ok",
  "output": [{ "summary": {...}, "items": [...] }],
  "requiresApproval": null
}

Handle approval

If the workflow needs approval:

{
  "status": "needs_approval",
  "output": [],
  "requiresApproval": {
    "prompt": "Send 3 draft replies?",
    "items": [...],
    "resumeToken": "..."
  }
}

Present the prompt to the user. If they approve:

{
  "action": "resume",
  "token": "<resumeToken>",
  "approve": true
}

Example workflows

Email triage

gog.gmail.search --query 'newer_than:1d' --max 20 | email.triage

Fetches recent emails, classifies into buckets (needs_reply, needs_action, fyi).

Email triage with approval gate

gog.gmail.search --query 'newer_than:1d' | email.triage | approve --prompt 'Process these?'

Same as above, but halts for approval before returning.

Key behaviors

Deterministic: Same input → same output (no LLM variance in pipeline execution)
Approval gates: approve command halts execution, returns token
Resumable: Use resume action with token to continue
Structured output: Always returns JSON envelope with protocolVersion