sprintx-openclaw-handoff

SprintX OpenClaw Handoff

Standalone public repo for SprintX's first ClawHub skill.

This skill helps an existing OpenClaw operator connect SprintX with the minimum proof packet that SprintX can verify today:

sx auth
-> sx project use
-> sx event
-> sx artifact add
-> sx status
-> sx brief

It is intentionally narrow.

• No custom MCP server
• No hidden SprintX automation
• No generic natural-language control plane
• No project creation flow inside the skill

Links:

• ClawHub: https://clawhub.ai/terry3838/sprintx-openclaw-handoff
• GitHub: https://github.com/terry3838/SprintX-SKILL
• SprintX handoff quickstart: https://www.sprintx.co.kr/docs/getting-started/openclaw-handoff-quickstart
• SprintX CLI quickstart: https://www.sprintx.co.kr/docs/getting-started/cli-quickstart
• Korean README: README.ko.md
• Chinese README: README.zh.md

Why This Exists

OpenClaw users can already connect to SprintX with the shipped CLI.

The problem is first-use friction. The first connection still asks people to remember too much:

• which command comes first
• where the projectId comes from
• what success looks like
• what to do when the packet fails halfway through

This skill compresses that into one operator-facing flow.

Not more magic. Less confusion.

What The Skill Does

The skill guides the operator through:

1. Checking SprintX prerequisites
2. Verifying that sx already exists before recommending install
3. Using the handoff card or handoff URL as the primary projectId source
4. Falling back to sx project list only when recovery is needed
5. Running the proof packet in the correct order
6. Verifying success with sx status and sx brief
7. Handling common failures with explicit rescue guidance

Who This Is For

Use this if:

• you already use OpenClaw
• you already have a SprintX account
• you already have a SprintX project or can create one on the web
• you want the safest current handoff path, not a broad integration surface

Do not use this skill if you want:

• SprintX project creation from chat
• task CRUD from the skill itself
• review or approval workflows
• a remote MCP server

Quick Start

1. Install the skill

From ClawHub:

clawhub install sprintx-openclaw-handoff

If your environment requires non-interactive install and you have already reviewed the skill:

clawhub install sprintx-openclaw-handoff --force --no-input

2. Let the skill walk the operator

Typical prompt:

Use $sprintx-openclaw-handoff to connect my existing OpenClaw runtime to SprintX.

3. Follow the proof packet

The skill will keep the operator on the narrow path:

sx auth
sx project use <project-id>
sx event --type runtime.started --summary "OpenClaw handoff started"
sx artifact add --title "first-log" --reference-uri "file:///tmp/openclaw.log" --content-type "text/plain" --summary "Initial OpenClaw evidence"
sx status
sx brief

Prerequisites

Before the handoff starts, the operator should have:

• a SprintX web session
• a SprintX project
• access to /dashboard?handoff=1&projectId=<id> or the handoff card

If the projectId is already visible there, the skill uses it directly.

sx project list is not the default first step. It is recovery only.

Trust Boundary

This is important.

• SprintX is not the executor. OpenClaw executes, SprintX reads and governs.
• The default auth path is browser-approved sx auth.
• Access-token override is treated as advanced or break-glass only.
• The skill explicitly warns against pasting tokens into chat.
• Provider API keys are outside this flow.

Advanced Operators

Headless and CI-style users are supported, but not as the primary path.

Examples:

sx --headless auth

The skill also documents the advanced access-token override, but keeps it out of the main onboarding path on purpose.

Repo Structure

SprintX-SKILL/
├── SKILL.md
├── README.md
├── CHANGELOG.md
├── LICENSE
├── LICENSE.md
├── package.json
├── references/
│   └── source-of-truth.md
├── scripts/
│   ├── check-skill.mjs
│   └── check-contract.mjs
└── .github/workflows/ci.yml

Design choice:

• SKILL.md is the only operator-truth document
• README.md is for humans evaluating the repo
• references/ holds upstream links and source-of-truth docs
• CI validates the repo, but publish stays manual

Validation

Run:

npm run check

This verifies:

• frontmatter shape
• declared runtime requirements
• install metadata
• license presence
• command order
• recovery-only sx project list
• advanced-only token guidance

Release Model

This repo uses validation CI plus manual ClawHub publish.

Publish:

clawhub publish . \
  --slug sprintx-openclaw-handoff \
  --name "SprintX OpenClaw Handoff" \
  --version 0.1.3 \
  --tags latest \
  --changelog "Add localized READMEs and tighten metadata to match real requirements"

We keep publish manual on purpose. The current ClawHub docs clearly support CLI-driven skill publish, while plugin/package auto-publish flows are more mature than skill auto-publish flows today.

Note:

• In the locally verified clawhub CLI, the working command is top-level clawhub publish <path>.
• Some upstream docs still describe clawhub skill publish.
• Re-check the installed CLI help before cutting a release.

Korean Note

이 저장소는 SprintX의 첫 공개 OpenClaw skill 저장소입니다.

• 기본 경로는 sx auth -> sx project use -> sx event -> sx artifact add -> sx status -> sx brief
• projectId는 handoff URL/card가 우선
• sx project list는 복구 경로
• access-token override는 기본 경로가 아니라 advanced-only

한국어 상세 문서:

• https://www.sprintx.co.kr/docs/getting-started/openclaw-handoff-quickstart
• https://www.sprintx.co.kr/docs/getting-started/cli-quickstart.ko