← Back to Skills Marketplace
zoujiejun

Agent Todo

by zoujiejun · GitHub ↗ · v1.2.1 · MIT-0
cross-platform ⚠ suspicious
173
Downloads
1
Stars
1
Active Installs
8
Versions
Install in OpenClaw
/install agent-todo
Description
Turn follow-up promises into an execution queue for agents. Use when chats or discussions create tasks that should be claimed and executed during heartbeat,...
README (SKILL.md)

agent-todo Skill

Use this skill as an execution queue, not as a passive reminder list.

核心原则:每个 agent 只维护并消费自己 workspace 下的任务队列;需要跨 agent 分发时,再按 OpenClaw 已注册的 workspace 进行发现和投递。

Core commands

bash ./script.sh add "Publish release" \
  --task-type publish \
  --source "forum:#19/reply:88" \
  --next-action "Push main to GitHub and publish ClawHub version" \
  --success-criteria "GitHub and ClawHub are both updated"

bash ./script.sh dispatch "Review release" \
  --to-agent reviewer \
  --task-type review \
  --source "chat:direct" \
  --next-action "Review release artifacts" \
  --success-criteria "Feedback delivered"

bash ./script.sh run-pending --claim
bash ./script.sh done \x3Cid> --note "what was completed"
bash ./script.sh report \x3Cid>
bash ./script.sh block \x3Cid> --reason "why blocked"
bash ./script.sh setup-heartbeat --write
bash ./script.sh setup-heartbeat --all --write

Workflow

  1. Add tasks with enough execution context:
    • task_type
    • next_action
    • success_criteria
    • source
  2. For composite goals, prefer plan to split them into concrete steps.
  3. During heartbeat, run:
    • bash ./script.sh run-pending --claim
  4. If it returns EXECUTE_NOW, do the task immediately.
  5. Prefer continuing a running task before opening a fresh pending one.
  6. To assign work to another agent, use dispatch --to-agent \x3Cagent_id>.
  7. After execution:
    • success → done
    • generate reply text → report
    • cannot continue → block
    • no longer needed → cancel

Storage model

  • Current workspace queue: .agent-todo/tasks.json
  • Optional local identity: .agent-todo/local.json
  • Workspace discovery source: ~/.openclaw/openclaw.json
  • Heartbeat wiring: managed block in HEARTBEAT.md

Do not hand-write workspace paths in normal usage. Let the script resolve the current workspace and discover registered workspaces from OpenClaw.

Notes

  • Single-workspace mode works out of the box after install.
  • Multi-agent routing is opt-in: it only matters when you call dispatch.
  • setup-heartbeat --all --write appends or updates a managed block for every discovered workspace instead of overwriting the full file.
  • report generates different output shapes for forum sources and direct chat sources.
Usage Guidance
This skill appears to do exactly what it says: manage a local task queue and optionally dispatch tasks to other workspaces discovered from your OpenClaw config. Before installing: (1) ensure python3 and a POSIX shell are available (scripts assume them); (2) review and back up any HEARTBEAT.md files you care about because setup-heartbeat --write will insert/update a managed block; (3) enable the post_reply hook only if you want automatic conversion of forum/chat replies into queued tasks (its heuristics may false-positive); (4) understand that dispatch will write to other workspaces discovered via ~/.openclaw/openclaw.json — ensure that config and workspace permissions are appropriate. If you want higher assurance, inspect the full scripts locally (they are included) and run the provided smoke tests in a temporary workspace first.
Capability Analysis
Type: OpenClaw Skill Name: agent-todo Version: 1.2.1 The skill implements a task execution queue that integrates with the OpenClaw heartbeat system. It possesses high-risk capabilities, including the ability to modify 'HEARTBEAT.md' files for persistence and read the global '~/.openclaw/openclaw.json' configuration to discover and write to other agent workspaces (scripts/agent_todo.py). While these behaviors are aligned with the stated purpose of multi-agent task coordination and automated execution, they represent a significant attack surface. Additionally, the 'post_reply.sh' hook automatically queues tasks based on agent replies, and the 'SKILL.md' instructions direct the agent to execute tasks 'immediately' upon discovery, which could be exploited if an agent is manipulated into committing to malicious tasks. No evidence of intentional malice was found.
Capability Assessment
Purpose & Capability
Name/description match the implementation: scripts create and manage a local tasks.json, claim work during heartbeat, and optionally dispatch tasks to discovered workspaces. The only minor mismatch is that the package metadata declares no required binaries while the shipped scripts assume a POSIX shell and python3 are available.
Instruction Scope
Runtime instructions and hooks operate on workspace-local files (.agent-todo/tasks.json, .agent-todo/local.json, HEARTBEAT.md) and on OpenClaw's config (~/.openclaw/openclaw.json). Hooks can auto-create tasks from reply content and will call the CLI to add/dispatch tasks. This matches the stated purpose, but users should note enabling the post_reply hook will automatically queue tasks when reply content matches the heuristic.
Install Mechanism
No remote install or downloads are declared; this is an instruction+script bundle. Nothing is fetched from external URLs and no archives are extracted, which lowers installation risk.
Credentials
No secrets or unrelated environment variables are required. The code optionally reads OPENCLAW_CONFIG, AGENT_TODO_WORKSPACE, and TODO_DB to locate workspaces — these are proportional to discovery and workspace binding. It also relies on standard shell/python runtime environment (not declared in metadata).
Persistence & Privilege
The skill writes and updates local workspace state (.agent-todo/*) and can modify HEARTBEAT.md (append/update a managed block). When dispatch or setup-heartbeat --all are used, it will write into other discovered workspaces; this is expected for cross-agent routing but is a persistence action users should be aware of.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install agent-todo
  3. After installation, invoke the skill by name or use /agent-todo
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.2.1
Fix heartbeat workspace binding for shared/external installs; add smoke coverage.
v1.1.5
Refresh README opening and AI SEO description
v1.1.4
Align ClawHub display name with Agent Forum style
v1.1.3
Repair TODO schema drift and validate completed task rows
v1.1.2
Use bash-prefixed commands to avoid executable-bit issues after install
v1.1.1
Add heartbeat onboarding commands and cross-links between GitHub and ClawHub
v1.1.0
Execution queue release with planning, source-aware reports, and smoke test
v1.0.0
Initial public release
Metadata
Slug agent-todo
Version 1.2.1
License MIT-0
All-time Installs 1
Active Installs 1
Total Versions 8
Frequently Asked Questions

What is Agent Todo?

Turn follow-up promises into an execution queue for agents. Use when chats or discussions create tasks that should be claimed and executed during heartbeat,... It is an AI Agent Skill for Claude Code / OpenClaw, with 173 downloads so far.

How do I install Agent Todo?

Run "/install agent-todo" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is Agent Todo free?

Yes, Agent Todo is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does Agent Todo support?

Agent Todo is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created Agent Todo?

It is built and maintained by zoujiejun (@zoujiejun); the current version is v1.2.1.

💬 Comments