← Back to Skills Marketplace
2915
Downloads
1
Stars
6
Active Installs
1
Versions
Install in OpenClaw
/install agent-task-manager
Description
Manages and orchestrates multi-step, stateful agent workflows; handles task dependencies, persistent state, error recovery, and external rate-limiting. Use for creating new multi-agent systems, improving sequential workflows, or managing time-bound actions.
README (SKILL.md)
Agent Task Manager
Overview
This skill provides the structure and primitives for building resilient, complex, and professional multi-agent systems within the OpenClaw environment. It transforms simple scripts into production-ready workflows.
Core Capabilities
1. Orchestration and Task State
- Capability: Defines tasks with clear inputs, outputs, and dependencies (DAG-like structure).
- Execution: Uses
molt_task.pyto manage state intask_state.json. - Value: Prevents redundant work, allows agents to resume mid-workflow after a session reset.
2. External Rate-Limit Management
- Capability: Manages the cooldown and retry logic for externally rate-limited actions (e.g., API posts, web scrapes).
- Execution: Uses the
scripts/cooldown.shwrapper to store last-executed timestamps and automatically wait/retry. - Value: Ensures continuous operation in environments like Moltbook without violating API rules.
3. Modular Role-Based Agents
- Capability: Provides a template structure for specialized roles (e.g.,
ContractAuditor,FinancialAnalyst). - Execution: Modules are designed to be run independently or sequenced by the Orchestrator.
- Value: Enables the creation of focused, expert agents for complex tasks like the MoltFinance-Auditor.
Example Workflow: MoltFinance-Auditor
- Task:
FinancialAudit - Dependencies:
- Role 1:
ContractAuditor(Input: Contract Address, Output: Contract Safety Score) - Role 2:
FinancialAnalyst(Input: Contract Address + Safety Score, Output: Trust Score)
- Role 1:
- External Action:
MoltbookPost(Dependent on final Trust Score; subject to Rate Limit).
Resources
scripts/
molt_task.py: Python class for task state management.cooldown.sh: Shell wrapper for managing rate-limited executions.
references/
workflow_schema.md: JSON schema for defining complex task dependencies.rate_limit_patterns.md: Guide to handling common API rate limits (e.g., Moltbook, Helius).
Usage Guidance
This skill appears to do what it says: local orchestration, state persistence, and cooldown management. Before installing or running it, consider: 1) The cooldown script uses eval to run commands — if you feed untrusted strings into the wrapper or tasks, an attacker could run arbitrary shell commands. Replace eval with safer execution (array-style exec) or strictly validate/whitelist commands. 2) The skill writes task_state.json and agent_task_manager_data/ in the current working directory — run it in a sandboxed directory and review or back up those files. 3) Notification and external integrations are placeholders; if you wire in Signal/Moltbook/other APIs you will need to supply credentials — ensure you store tokens securely and only grant least privilege. 4) Audit any tasks created from user input to ensure they cannot inject shell commands or arbitrary actions. If you plan to run in production or give the skill network access, review and harden the code (remove eval, validate inputs, and implement explicit API clients) and consider running it in an isolated environment first.
Capability Analysis
Type: OpenClaw Skill
Name: agent-task-manager
Version: 1.0.0
The `scripts/cooldown.sh` file uses `eval "$COMMAND"`, which allows for arbitrary command execution. While intended for a benign purpose (rate-limiting a wrapped command), this is a high-risk primitive that could be exploited if the `$COMMAND` argument is supplied by untrusted input, potentially leading to unauthorized code execution by the OpenClaw agent. Additionally, the `orchestrator.py` and `task_parser.py` define and process task structures with `role`, `action`, and `message` fields, which, if not properly sanitized or if the underlying tools are vulnerable, could be used for prompt injection or command execution, although the current implementation only simulates these actions.
Capability Assessment
Purpose & Capability
Name/description align with the provided files: orchestrator, task state manager, parser, and a cooldown wrapper implement the described multi-step workflows, state persistence, and rate-limit handling. The example workflows and role templates are consistent with the code (roles are simulated/placeholders).
Instruction Scope
SKILL.md and the scripts remain within the orchestration domain, but cooldown.sh executes an arbitrary COMMAND with eval("$COMMAND") — this permits arbitrary shell execution if untrusted input reaches it. The orchestrator currently uses simulated role functions (no network calls), but the code writes/reads local state files (task_state.json, agent_task_manager_data/*).
Install Mechanism
No install spec — instruction-only with bundled scripts. Nothing is downloaded or installed automatically, so no install-time network code to review.
Credentials
The skill declares no required credentials or env vars, and code does not access external tokens. However example workflows reference external channels (Signal, Moltbook) and the orchestrator contains placeholders for sending messages; integrating real notification or API calls will require adding credentials later. Currently there is no unexplained credential access.
Persistence & Privilege
The skill persists state locally (task_state.json) and creates a timestamp directory agent_task_manager_data/ for cooldowns. It does not request always:true, does not modify other skills, nor attempt system-wide config changes.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install agent-task-manager - After installation, invoke the skill by name or use
/agent-task-manager - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Initial launch of the multi-agent orchestration and rate-limit management skill.
Metadata
Frequently Asked Questions
What is Agent Task Manager?
Manages and orchestrates multi-step, stateful agent workflows; handles task dependencies, persistent state, error recovery, and external rate-limiting. Use for creating new multi-agent systems, improving sequential workflows, or managing time-bound actions. It is an AI Agent Skill for Claude Code / OpenClaw, with 2915 downloads so far.
How do I install Agent Task Manager?
Run "/install agent-task-manager" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Agent Task Manager free?
Yes, Agent Task Manager is completely free (open-source). You can download, install and use it at no cost.
Which platforms does Agent Task Manager support?
Agent Task Manager is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Agent Task Manager?
It is built and maintained by dobbybud (@dobbybud); the current version is v1.0.0.
More Skills