← 返回 Skills 市场
uchibeke

Aport Handoff

作者 Uchi Uchibeke · GitHub ↗ · v0.1.0 · MIT-0
cross-platform ✓ 安全检测通过
207
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install aport-handoff
功能描述
Package completed work for handoff to another agent or a human. Every item in the handoff must have a verified APort decision. The handoff document is verifi...
使用说明 (SKILL.md)

/aport-handoff — Verified Work Package

A handoff is a promise. This skill makes that promise verifiable. Every item in an APort handoff is backed by a signed APort decision. The recipient — human or agent — can independently verify every claim.

When to use this skill

  • When passing work to another agent in a pipeline
  • When completing your part of a multi-agent workflow
  • When handing a feature to a human for final review
  • When a sprint or milestone is complete and needs sign-off
  • When onboarding a new agent to a codebase you've been working on

Prerequisites

You must have an APort passport and completed tasks with ALLOW decisions. Tasks without verified decisions cannot be included in a verified handoff.

If you don't have a passport:

Step 1 — Identify what to hand off

Ask the user or determine from context:

  • What is the scope? (PR, feature, sprint, project)
  • Who is the recipient? (human, another agent, a team)
  • What time window? (today, this week, this sprint)

Step 2 — Fetch verified decisions for the scope

GET https://aport.io/api/verify/decisions/YOUR_AGENT_ID

Filter to decisions where allow: true within the relevant time window. Only ALLOW decisions are included in a handoff. DENY decisions indicate incomplete work — surface these separately as blockers.

Step 3 — Build the handoff document

# Handoff — [SCOPE] — [DATE]
**From:** [YOUR AGENT NAME] (aport.id/passport/YOUR_SLUG)
**To:** [RECIPIENT]
**APort verification:** All items below have signed decisions

---

## Completed Work

### [task description or title]
**Summary:** [from verification context]
**Output type:** [code / document / analysis / plan]
**Completed:** [decision created_at]
**Decision:** [decision_id] — verify at https://aport.io/api/verify/decisions/get/[decision_id]

Evidence provided:
- [criterion 1]: [evidence string from attestation]
- [criterion 2]: [evidence string from attestation]

---

[repeat for each completed item]

---

## Not Included (Incomplete)

The following tasks were attempted but not verified complete:

| Task | Last attempt | Blocked on |
|------|-------------|------------|
| [description] | [time] | [reason code from deny decision] |

These require attention before they can be handed off.

---

## What the recipient needs to know

[Contextual notes — dependencies, known issues, next steps, decisions made]
[Write this section yourself based on what the recipient needs]

---

## Verification

All decisions in this handoff are cryptographically signed by APort.
Verify any decision:
  GET https://aport.io/api/verify/decisions/get/DECISION_ID

Issuing agent passport:
  https://aport.id/passport/YOUR_SLUG

Step 4 — Deliver the handoff

Offer to deliver the handoff document through available channels:

  • Save to a file (e.g. handoff-[date].md)
  • Open a GitHub issue or PR description
  • Post to Slack or Discord (with user permission)
  • Send via any messaging tool available as an MCP tool

Always save a local copy regardless of delivery method.

What makes a handoff invalid

A handoff is invalid if:

  • It includes tasks without APort decisions
  • It claims completion for tasks still showing DENY
  • The decision_ids cannot be verified at the APort API
  • The issuing agent's passport is suspended

If any of these conditions apply, surface them clearly rather than including unverified items in the handoff.

Tone

Handoffs are functional documents. No padding. State what was done, what wasn't done, and what the recipient needs to know. The decision trail is the credibility — the prose is just navigation.

Links

安全使用建议
This skill is coherent and lightweight, but note it will perform network verification against aport.io and produce a local handoff file. Before installing or invoking it: 1) confirm aport.id / aport.io are services you trust and that decision IDs you publish are safe to share; 2) be explicit about where the agent may save files and whether it may post to GitHub/Slack/Discord — grant those permissions only when you approve; 3) if your APort decisions are private or require authentication, verify how your agent will authenticate (the skill doesn't request credentials itself); and 4) if you do not want any external network calls, do not allow the agent to run this skill or block network access for it.
功能分析
Type: OpenClaw Skill Name: aport-handoff Version: 0.1.0 The skill is a utility designed to generate handoff documentation by fetching task verification data from the APort API (aport.io). It guides the agent through collecting 'verified decisions' via HTTP GET requests and formatting them into a markdown report for human or agent review. No evidence of data exfiltration, malicious execution, or prompt injection was found; the external links and CLI suggestions (npx aport-id) are consistent with the stated purpose of the APort verification service.
能力评估
Purpose & Capability
The skill's name/description (packaging verified work for handoff) matches the actions it requires: fetching APort decisions and assembling a handoff document. It does not request unrelated credentials, binaries, or config paths.
Instruction Scope
SKILL.md instructs the agent to perform HTTP GETs to aport.io verification endpoints, assemble a markdown handoff, save a local copy, and optionally deliver it via GitHub, Slack, Discord, or other messaging tools (with user permission). These actions are within the stated scope, but they involve network calls and writing a local file — the user should be aware the agent will contact an external API and may post to third-party channels if allowed.
Install Mechanism
No install spec and no code files are present (instruction-only), so nothing is written to disk at install time and no third-party packages are pulled in.
Credentials
The skill declares no required environment variables or credentials. It mentions using an APort passport and optional CLI (npx aport-id) as a prerequisite, which is reasonable for verifying decisions and is not requested as a secret by the skill itself.
Persistence & Privilege
always:false and no special system-level privileges are requested. The skill does ask the agent to save a local copy of the handoff and optionally send it to external channels, but it does not modify other skills or request persistent elevated privileges.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install aport-handoff
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /aport-handoff 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v0.1.0
Initial release of aport-handoff skill - Enables packaging and delivering completed work for handoff with verifiable APort decisions. - Builds a handoff document containing proof-backed records of completed tasks, referencing signed APort decisions. - Clearly surfaces incomplete or blocked tasks with required attention. - Provides guidance for document delivery and verification best practices. - Ensures recipients can independently verify every work claim via APort.
元数据
Slug aport-handoff
版本 0.1.0
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 1
常见问题

Aport Handoff 是什么?

Package completed work for handoff to another agent or a human. Every item in the handoff must have a verified APort decision. The handoff document is verifi... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 207 次。

如何安装 Aport Handoff?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install aport-handoff」即可一键安装,无需额外配置。

Aport Handoff 是免费的吗?

是的,Aport Handoff 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

Aport Handoff 支持哪些平台?

Aport Handoff 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 Aport Handoff?

由 Uchi Uchibeke(@uchibeke)开发并维护,当前版本 v0.1.0。

💬 留言讨论