hitl-workflow — community hitl-workflow, Personal-AI-Employee, community, ide skills, Claude Code, Cursor, Windsurf

v1.0.0

Об этом навыке

Идеально для ИИ-агентов, которым требуются безопасные рабочие процессы утверждения с человеческим участием с помощью Python Watchers и хранилищ памяти на основе Obsidian. A framework for building personal AI employees. Features a Sense-Think-Act loop using Python Watchers, Claude Code reasoning, and an Obsidian-based persistent memory vault.

AbdulSamad94 AbdulSamad94
[1]
[0]
Updated: 3/15/2026

Killer-Skills Review

Decision support comes first. Repository text comes second.

Reference-Only Page Review Score: 7/11

This page remains useful for operators, but Killer-Skills treats it as reference material instead of a primary organic landing page.

Original recommendation layer Concrete use-case guidance Explicit limitations and caution
Review Score
7/11
Quality Score
31
Canonical Locale
en
Detected Body Locale
en

Идеально для ИИ-агентов, которым требуются безопасные рабочие процессы утверждения с человеческим участием с помощью Python Watchers и хранилищ памяти на основе Obsidian. A framework for building personal AI employees. Features a Sense-Think-Act loop using Python Watchers, Claude Code reasoning, and an Obsidian-based persistent memory vault.

Зачем использовать этот навык

Наделяет агентов возможностью работать в рамках прозрачной структуры принятия решений с использованием рассуждений Claude Code, гарантируя, что каждое действие, влияющее на внешний мир, требует явного одобрения человека через хранилища постоянной памяти на основе Obsidian и Python Watchers.

Подходит лучше всего

Идеально для ИИ-агентов, которым требуются безопасные рабочие процессы утверждения с человеческим участием с помощью Python Watchers и хранилищ памяти на основе Obsidian.

Реализуемые кейсы использования for hitl-workflow

Реализация утверждения человека в цикле для критических процессов принятия решений
Использование Python Watchers для мониторинга в реальном времени и запросов на утверждение
Управление безопасными и прозрачными процессами принятия решений с хранилищами памяти на основе Obsidian

! Безопасность и ограничения

  • Требует явного одобрения человека для каждого внешнего действия
  • Зависит от Python Watchers и рассуждений Claude Code для работы
  • Требует хранилищ постоянной памяти на основе Obsidian для безопасного принятия решений

Why this page is reference-only

  • - Current locale does not satisfy the locale-governance contract.
  • - The underlying skill quality score is below the review floor.

Source Boundary

The section below is imported from the upstream repository and should be treated as secondary evidence. Use the Killer-Skills review above as the primary layer for fit, risk, and installation decisions.

After The Review

Decide The Next Action Before You Keep Reading Repository Material

Killer-Skills should not stop at opening repository instructions. It should help you decide whether to install this skill, when to cross-check against trusted collections, and when to move into workflow rollout.

Labs Demo

Browser Sandbox Environment

⚡️ Ready to unleash?

Experience this Agent in a zero-setup browser environment powered by WebContainers. No installation required.

Boot Container Sandbox

FAQ & Installation Steps

These questions and steps mirror the structured data on this page for better search understanding.

? Frequently Asked Questions

What is hitl-workflow?

Идеально для ИИ-агентов, которым требуются безопасные рабочие процессы утверждения с человеческим участием с помощью Python Watchers и хранилищ памяти на основе Obsidian. A framework for building personal AI employees. Features a Sense-Think-Act loop using Python Watchers, Claude Code reasoning, and an Obsidian-based persistent memory vault.

How do I install hitl-workflow?

Run the command: npx killer-skills add AbdulSamad94/Personal-AI-Employee. It works with Cursor, Windsurf, VS Code, Claude Code, and 19+ other IDEs.

What are the use cases for hitl-workflow?

Key use cases include: Реализация утверждения человека в цикле для критических процессов принятия решений, Использование Python Watchers для мониторинга в реальном времени и запросов на утверждение, Управление безопасными и прозрачными процессами принятия решений с хранилищами памяти на основе Obsidian.

Which IDEs are compatible with hitl-workflow?

This skill is compatible with Cursor, Windsurf, VS Code, Trae, Claude Code, OpenClaw, Aider, Codex, OpenCode, Goose, Cline, Roo Code, Kiro, Augment Code, Continue, GitHub Copilot, Sourcegraph Cody, and Amazon Q Developer. Use the Killer-Skills CLI for universal one-command installation.

Are there any limitations for hitl-workflow?

Требует явного одобрения человека для каждого внешнего действия. Зависит от Python Watchers и рассуждений Claude Code для работы. Требует хранилищ постоянной памяти на основе Obsidian для безопасного принятия решений.

How To Install

  1. 1. Open your terminal

    Open the terminal or command line in your project directory.

  2. 2. Run the install command

    Run: npx killer-skills add AbdulSamad94/Personal-AI-Employee. The CLI will automatically detect your IDE or AI agent and configure the skill.

  3. 3. Start using the skill

    The skill is now active. Your AI agent can use hitl-workflow immediately in the current project.

! Reference-Only Mode

This page remains useful for installation and reference, but Killer-Skills no longer treats it as a primary indexable landing page. Read the review above before relying on the upstream repository instructions.

Upstream Repository Material

The section below is imported from the upstream repository and should be treated as secondary evidence. Use the Killer-Skills review above as the primary layer for fit, risk, and installation decisions.

Upstream Source

hitl-workflow

Install hitl-workflow, an AI agent skill for AI agent workflows and automation. Works with Claude Code, Cursor, and Windsurf with one-command setup.

SKILL.md
Readonly
Upstream Repository Material
The section below is imported from the upstream repository and should be treated as secondary evidence. Use the Killer-Skills review above as the primary layer for fit, risk, and installation decisions.
Supporting Evidence

Skill: Human-in-the-Loop (HITL) Approval Workflow

Core Principle

You are a powerful AI agent but you are NOT autonomous by default. Every action that affects the outside world requires explicit human approval. This is not optional. This is the most important rule you follow.

The Three Folders

FolderPurposeWho Acts
vault/Pending_Approval/You write approval requests hereYou (Claude)
vault/Approved/Human moves files here to approveHuman
vault/Rejected/Human moves files here to rejectHuman

You WRITE to Pending_Approval/. You NEVER write directly to Approved/. You CHECK Approved/ to know when to execute. You NEVER act without checking Approved/ first.

When HITL is REQUIRED (No Exceptions)

  • Sending any email
  • Posting to any social media (LinkedIn, Twitter, Instagram)
  • Any action involving money or payments
  • Replying to any client or business contact
  • Deleting or modifying any external data
  • Any action that cannot be easily undone
  • Any action involving a new contact with no prior history
  • Any action that involves making a commitment on your behalf

When HITL is NOT Required

  • Reading files
  • Creating plan files
  • Updating Dashboard.md
  • Moving files between internal vault folders
  • Writing logs
  • Organizing and renaming internal files

Approval File Template

For ALL external actions, create this file in vault/Pending_Approval/:

Filename format: [ACTION_TYPE][description][YYYYMMDD_HHMMSS].md


type: approval_request action: [email_send / linkedin_post / payment / whatsapp_reply / other] to: [recipient — email, phone, or platform] subject: [subject if applicable] amount: [dollar amount if payment] task_source: [original task filename that triggered this] created: [ISO timestamp] expires: [ISO timestamp — 24 hours for email, 48 hours for posts, 1 hour for payments] priority: [normal / urgent / critical] status: pending


What Will Happen If You Approve

[Describe in plain language exactly what action will be taken]

Content / Details

[The exact content — email body, post text, payment details, etc.]

Why This Action

[Brief explanation of why this needs to happen]

Risks If Not Approved

[What happens if this is rejected or ignored]

✅ To Approve

Move this file to vault/Approved/

✏️ To Edit Then Approve

Edit the content above, then move to vault/Approved/

❌ To Reject

Move this file to vault/Rejected/

⏰ Expiry

This approval expires at [expiry timestamp]. After expiry, a new approval request must be created.

Checking for Approvals

When you need to execute pending actions, follow this process:

  1. List all files in vault/Approved/
  2. For each approved file: a. Read the file b. Verify it hasn't expired (check the expires field) c. Verify the action type matches what you're about to do d. Execute the action via the appropriate MCP server e. Log the execution f. Move the file from vault/Approved/ to vault/Done/
  3. If no files in vault/Approved/, wait — do not act

Handling Rejections

When a file appears in vault/Rejected/:

  1. Read it to understand what was rejected
  2. Check if there are any notes added by the human
  3. Update the original task file status to: rejected
  4. Log: [timestamp] | rejected | [filename] | [action_type]
  5. Move the rejection file to vault/Done/
  6. Do NOT retry the action unless explicitly asked

Handling Expired Approvals

If an approval file exists in vault/Pending_Approval/ past its expiry:

  1. Move it to vault/Rejected/ with a note: "Expired — no action taken"
  2. Log the expiry
  3. Create a new task in vault/Needs_Action/ if the action is still needed: FILE: REPROCESS_[original_task]_[timestamp].md Content: "Previous approval expired. Human review needed to determine if this should be re-processed."

Payment-Specific Rules

Payments have stricter rules than all other actions:

  • Expiry is 1 hour (not 24)
  • Require the exact recipient name, bank details, and reference
  • NEVER retry a payment automatically even if it failed
  • NEVER send payment to a new recipient without prior history
  • Flag any payment over $500 as CRITICAL priority
  • Always include the invoice or reference document

Связанные навыки

Looking for an alternative to hitl-workflow or another community skill for your workflow? Explore these related open-source skills.

Показать все

openclaw-release-maintainer

Logo of openclaw
openclaw

Your own personal AI assistant. Any OS. Any Platform. The lobster way. 🦞

widget-generator

Logo of f
f

Создание настраиваемых плагинов виджетов для системы ленты новостей prompts.chat

flags

Logo of vercel
vercel

Фреймворк React

138.4k
0
Браузер

pr-review

Logo of pytorch
pytorch

Tensors and Dynamic neural networks in Python with strong GPU acceleration

98.6k
0
Разработчик