Task Management logo

Task Management

Community
chenhg5
task-management

Create, assign, monitor, and close tasks across agents. Covers full task lifecycle — add, priority, dependencies, confirm-request, retry, cancel, bulk ops, and cross-agent search.

Overview

Publisherchenhg5
Repositoryagencycli
Skill nametask-management
Stars
240
Forks
26
Bundled files
Instructions only
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by chenhg5 on GitHub. Read the source before you install it.

Installation

Install the Task Management AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/chenhg5/agencycli/tree/main/skills/task-management
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/chenhg5/agencycli.git /tmp/agencycli
mkdir -p .claude/skills
cp -r /tmp/agencycli/skills/task-management .claude/skills/task-management
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Task Management in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Task Management on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Task Management is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Skill: Task Management

Use this skill when your role involves planning, delegating, and tracking work across agents. The agency workspace is at $AGENCY_DIR.


Priority levels

ValueLabelWhen to use
0criticalBlocking production / blocking other agents
1highShould be picked up in the current cycle
2normalDefault — regular backlog work
3lowNice-to-have, do when free

Create a task for an agent

bash
# Minimal
agencycli --dir $AGENCY_DIR task add \
  --project <project> --agent <agent> \
  --title "Short title" \
  --prompt "Detailed instructions for the agent..."

# With priority and type
agencycli --dir $AGENCY_DIR task add \
  --project <project> --agent <agent> \
  --title "Fix auth bug" \
  --prompt "Reproduce with: curl -X POST /login with empty password. Root cause is in auth/validator.go." \
  --priority 1 \
  --type bug

# With dependencies (task will not start until listed IDs are done)
agencycli --dir $AGENCY_DIR task add \
  --project <project> --agent <agent> \
  --title "Deploy to staging" \
  --prompt "Run: make deploy-staging" \
  --depends-on <task-id-1> --depends-on <task-id-2>

# With scheduling / nesting metadata
agencycli --dir $AGENCY_DIR task add \
  --project <project> --agent <agent> \
  --title "Implement sub-feature" \
  --prompt "..." \
  --parent <parent-task-id> \
  --due-date 2026-07-15 \
  --estimate-duration 2h

Task metadata fields

FieldCLI flagNotes
Due date--due-date YYYY-MM-DDDeadline
Estimate--estimate-duration 30mGo duration (30m, 2h)
Parent--parent <task-id>Sub-task nesting
Labels--label (repeatable)Tags
Started / finished(auto)Set when status → in_progress / terminal

Task types

feature · bug · chore · review · deploy · research · wakeup · (custom string)


Inspect task queues

bash
# All active tasks for one agent (sorted by priority)
agencycli --dir $AGENCY_DIR task list --project <project> --agent <agent>

# Filter by status
agencycli --dir $AGENCY_DIR task list --project <project> --agent <agent> --status pending
agencycli --dir $AGENCY_DIR task list --project <project> --agent <agent> --status in_progress

# Include archived (completed / cancelled) tasks
agencycli --dir $AGENCY_DIR task list --project <project> --agent <agent> --archived

# Show full detail of a specific task (project + agent known)
agencycli --dir $AGENCY_DIR task show <task-id> --project <project> --agent <agent>

# Find a task anywhere by ID (no project/agent needed)
agencycli --dir $AGENCY_DIR task find --id <task-id>

Update task fields

bash
# Only flags you pass are changed; project/agent auto-detected if omitted
agencycli --dir $AGENCY_DIR task set <task-id> --priority 1
agencycli --dir $AGENCY_DIR task set <task-id> --status in_progress
agencycli --dir $AGENCY_DIR task set <task-id> \
  --due-date 2026-07-15 --estimate-duration 2h --parent <parent-task-id>
agencycli --dir $AGENCY_DIR task set <task-id> --label bug --label urgent
agencycli --dir $AGENCY_DIR task set <task-id> --due-date ""   # clear due date
agencycli --dir $AGENCY_DIR task set <task-id> --format json   # print updated task

Updatable: title, description, status, priority, type, summary, label, parent, due-date, estimate-duration, assignee, prompt / prompt-file, position.

Status changes auto-maintain started_at / finished_at.


Monitor progress across agents

To get an overview of all agents and their queue depths:

bash
agencycli --dir $AGENCY_DIR overview

To iterate agents and check queues programmatically:

bash
# List all agents in a project
agencycli --dir $AGENCY_DIR list agents --project <project>

# Then per agent
agencycli --dir $AGENCY_DIR task list --project <project> --agent <agent> --status pending
agencycli --dir $AGENCY_DIR task list --project <project> --agent <agent> --status in_progress

Task stats (throughput & efficiency)

bash
# Today by executor queue (default: all agents, grouped)
agencycli --dir $AGENCY_DIR task stats --since today

# One agent's queue
agencycli --dir $AGENCY_DIR task stats --since today --project <project> --agent <agent>

# By assignee field
agencycli --dir $AGENCY_DIR task stats --since today --assignee <project>/<agent>
agencycli --dir $AGENCY_DIR task stats --since 7d --by assignee

# By label (evening summary: value / category buckets)
agencycli --dir $AGENCY_DIR task stats --since today --by label:value
agencycli --dir $AGENCY_DIR task stats --since today --by label:category
agencycli --dir $AGENCY_DIR task stats --since today --label value:owner

# List each finished task
agencycli --dir $AGENCY_DIR task stats --since today --detail

# JSON export (includes elapsedHuman, estimateHuman, etc.)
agencycli --dir $AGENCY_DIR task stats --since today --format json

Metrics: success/failed/cancelled counts, sum of actual elapsed time (started_at→finished_at), sum of estimate_duration, coverage (tasks with estimates), and actual/estimate ratio. JSON uses human-readable durations (elapsedHuman: 14m32s) instead of nanosecond integers.


bash
# Success
agencycli --dir $AGENCY_DIR task done \
  --id <task-id> --status success --summary "What was accomplished"

# Failure
agencycli --dir $AGENCY_DIR task done \
  --id <task-id> --status failed --error "reason"

Request human confirmation (non-blocking)

When an agent needs a human decision before continuing, it calls:

bash
agencycli --dir $AGENCY_DIR task confirm-request \
  --id $TASK_ID \
  --summary "One-line description of what needs approval" \
  --action-item "Option A: reply 'approve'" \
  --action-item "Option B: reply 'reject <reason>'"

The task is archived. The human responds via:

bash
agencycli --dir $AGENCY_DIR inbox list
agencycli --dir $AGENCY_DIR inbox messages    # view messages
agencycli --dir $AGENCY_DIR inbox reply <msg-id> --body "approve"
agencycli --dir $AGENCY_DIR inbox reject  <task-id> --reason "out of scope"

The human's reply is sent as a message. The agent sees it on the next wakeup and continues via session memory.


Escalate or delegate a task

If a task is better suited for a different agent, cancel it and re-create for the target agent:

bash
# 1. Cancel the original
agencycli --dir $AGENCY_DIR task cancel <task-id> --project <project> --agent <from-agent>

# 2. Create for the new agent with transferred context
agencycli --dir $AGENCY_DIR task add \
  --project <project> --agent <to-agent> \
  --title "<same title>" \
  --prompt "<original prompt + re-delegation note>"

To notify the new agent with async context before the task runs:

bash
agencycli --dir $AGENCY_DIR inbox send \
  --from <project>/pm \
  --to   <project>/<to-agent> \
  --subject "Incoming task: <title>" \
  --body "Extra context: ..."

Retry a failed task

bash
agencycli --dir $AGENCY_DIR task retry <task-id> \
  --project <project> --agent <agent>

Optionally update the prompt before retrying:

bash
agencycli --dir $AGENCY_DIR task retry <task-id> \
  --project <project> --agent <agent> \
  --prompt "Updated instructions based on failure: ..."

Cancel tasks

bash
# Cancel a specific task
agencycli --dir $AGENCY_DIR task cancel <task-id> \
  --project <project> --agent <agent>

# Emergency halt — cancel all pending tasks for all agents in a project
agencycli --dir $AGENCY_DIR task stop-all \
  --project <project> --all-agents

# Cancel pending + in-progress for a single agent
agencycli --dir $AGENCY_DIR task stop-all \
  --project <project> --agent <agent> --include-running

Recurring tasks with cron

Schedule a task to be automatically enqueued on a fixed schedule:

bash
agencycli --dir $AGENCY_DIR cron add \
  --project <project> --agent <agent> \
  --id weekly-review \
  --title "Weekly backlog review" \
  --schedule "0 9 * * 1" \
  --prompt "Review all open issues. Prioritise for the week. Update task queue accordingly."

agencycli --dir $AGENCY_DIR cron list    --project <project> --agent <agent>
agencycli --dir $AGENCY_DIR cron disable weekly-review --project <project> --agent <agent>
agencycli --dir $AGENCY_DIR cron enable  weekly-review --project <project> --agent <agent>

Cron syntax: minute hour day month weekday (standard 5-field)


Token usage and cost

bash
# One agent
agencycli --dir $AGENCY_DIR task tokens --project <project> --agent <agent>

# All agents in a project
agencycli --dir $AGENCY_DIR task tokens --project <project> --all-agents

# Specific task
agencycli --dir $AGENCY_DIR task tokens \
  --project <project> --agent <agent> --task <task-id>

Task assignment decision guide

SituationAction
Clear scope, right agenttask add directly
Scope unclear, need human input firstinbox send --to human to clarify, then task add
Depends on another task being done firsttask add --depends-on <id>
Task failed oncetask retry with updated prompt
Wrong agent received the taskCancel + re-create for correct agent
Repeated recurring workcron add with a schedule
Human must approve before agent proceedsAgent calls task confirm-request
Need to change title/priority/due date/parent mid-flighttask set <id> --<field> ...
Review throughput / efficiency for a person or agenttask stats --since today [--assignee A]
Evening summary by value/category labelstask stats --since today --by label:value or --by label:category
Need to know which agent owns a task IDtask find --id <id>

Frequently asked questions

What does the Task Management AI skill do?

Create, assign, monitor, and close tasks across agents. Covers full task lifecycle — add, priority, dependencies, confirm-request, retry, cancel, bulk ops, and cross-agent search.

Why use Task Management on TypingMind?

Because you install it once and use it with any model. Task Management is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Task Management in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/chenhg5/agencycli/tree/main/skills/task-management. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Task Management?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Task Management?

As many as you like. As long as a model supports skills, you can use Task Management with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Task Management AI skill free?

It is published on GitHub by chenhg5. Check the repository for licensing terms. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇