human-tasks

Create [HUMAN] tasks to route human-required workflow blockers.

88|108|Updated Apr 7, 2026
One-click install
npx skills add https://github.com/grandamenium/cortextos --skill human-tasks-grandamenium
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: human-tasks
Source: https://github.com/grandamenium/cortextos/tree/main/templates/analyst/.claude/skills/human-tasks
Command: npx skills add https://github.com/grandamenium/cortextos --skill human-tasks-grandamenium

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill prevents your agent workflow from stalling when the next step is impossible without a human making a decision, providing credentials, performing physical action, or completing an external manual task.

Core Features & Use Cases

  • Capability-block routing: Detects that a blocker is not a permission issue but a genuine human-required dependency and frames it as a [HUMAN] task.
  • Actionable, step-by-step instructions: Forces the task creator to write instructions specific enough that a human can complete the work without needing follow-up questions.
  • Operational unblocking behavior: Blocks the agent’s own task, notifies the orchestrator/user, and then resumes immediately once the human task is marked complete.

Quick Start

Create a new human task with a title starting with [HUMAN], assign it to human, put it in the human-tasks project, block your current task on it, and notify the orchestrator so it appears in the next briefing.

Frequently Asked Questions about human-tasks

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I handle workflow blockers that require human action in an automated agent pipeline?

Workflow blockers requiring human action are handled by creating explicitly labeled [HUMAN] tasks. The agent blocks its current task, notifies the orchestrator, and resumes immediately once the human completes the manual step.

What is human-in-the-loop task routing and when do I need it for my dashboard workflow?

Human-in-the-loop task routing separates genuine capability blockers from missing permissions. You need it when a workflow stalls on decisions, credential entry, or physical actions that only a human can complete.

How do I create step-by-step instructions for human tasks in an orchestration workflow?

Create a task titled starting with [HUMAN], assign it to a human, place it in the human-tasks project, and write actionable step-by-step instructions specific enough to complete without follow-up questions.

Can I use Telegram notifications to resume agent tasks after human approval?

Yes, notification triggers surface the [HUMAN] task in the dashboard and notify the orchestrator. The agent pipeline resumes immediately upon receiving the task completion notification.

What's the difference between a permission issue and a human capability blocker in task routing?

A permission issue is a missing access right, while a human capability blocker is a genuine dependency requiring manual decisions, payments, or physical actions. This routing frames the latter as a [HUMAN] task.

Why does my agent workflow stall on external manual tasks and how do I unblock it?

Agent workflows stall when the next step is impossible without human input. Unblock it by routing the dependency as a [HUMAN] task with correct assignee and project flags, blocking the agent until completion.