Agent-team-manager

Professional AI agent team management system for coordinating multiple OpenClaw agents. Implements the proven Otter Camp methodology with task assignment, pr...

Security Scan

The skill's code and runtime metadata contain multiple internal inconsistencies (missing main entry, undeclared/unused requirements) that make its behavior unclear; review and remediation are needed before trust or installation.

Purpose & Capability

Name/description (agent team management) matches the included JS modules (task allocation, tracking, quality, performance). However metadata and files diverge: package.json declares main: "index.js" but no index.js is present in the file manifest; example usage requires './index'. The manifest lists many JS modules but no entrypoint/exports, so it's unclear how the package is intended to be executed or integrated. Declared binary requirement 'python3' is not justified by the provided JavaScript source. These mismatches reduce confidence that required items are proportionate to the stated purpose.

Instruction Scope

SKILL.md describes local-only operation, Git integration, and workspace-based operation; its usage examples pass workspace paths explicitly. The SKILL.md's runtime instructions do not ask the agent to read unrelated system files or exfiltrate secrets. However the doc declares OPENCLAW_WORKSPACE as required env but the visible code does not reference process.env.OPENCLAW_WORKSPACE (no file in the included snippets uses that env), and example tasks mention analyzing an external QQ mailbox — an integration not implemented in the visible code. The instructions are generally scoped to team-management tasks but give the agent some latitude (e.g., 'works with Git') without concrete implementation details.

Install Mechanism

No install specification is provided (instruction-only), which is low risk, but the bundle actually contains multiple source files and a publish.sh. The presence of code + package.json without an install/build spec or entrypoint is an inconsistency. There is no remote download or extract step flagged. Overall installation behavior is unclear because required runtime binaries are declared but no install script uses them; that unpredictability is worth caution.

Credentials

Declared required env var: OPENCLAW_WORKSPACE (reasonable for a workspace-based manager). But the provided code snippets do not reference process.env.OPENCLAW_WORKSPACE or other environment variables, suggesting a mismatch between declared requirements and actual behavior. Declared required binaries include python3, but visible code is Node.js and does not call Python; git is listed (SKILL.md says it integrates with Git) but the code does not show git invocation. This disparity could indicate incomplete packaging or inaccurate metadata — both reduce transparency and raise risk because the real runtime needs aren't verifiable from the bundle.

Persistence & Privilege

Skill does not request always:true and does not claim to modify other skills or system-wide settings. No install spec attempts to create persistent system services. The code shown performs in-memory data management and logging; there are no obvious attempts to persist credentials or change agent configuration beyond its own structures.

What to consider before installing

Do not deploy this skill to production yet. The package has several internal inconsistencies that make its real behavior unclear: package.json points to index.js but index.js is missing; example code requires './index' but no entrypoint is present; the metadata lists python3 as a required binary although the code is JavaScript; OPENCLAW_WORKSPACE is declared required but isn’t referenced in the visible source. What to do before installing: 1) Ask the author/source for the missing index.js and a clear install/run procedure and confirm why python3 is required. 2) Inspect the missing/omitted files and any entrypoint to ensure there are no network calls or credential reads (search for process.env, http/https, net, child_process, fetch). 3) Run the code in an isolated sandbox/container and verify behavior (no unexpected outbound connections, no filesystem access outside the intended workspace). 4) Only provide workspace paths that do not expose sensitive data, and do not supply secrets or cloud credentials. 5) Prefer packages with a verifiable homepage, tests, and a clear entrypoint; if the author cannot provide the missing pieces and a clear explanation, treat the skill as untrusted.

AI Agent Team Manager

Overview

This skill implements a professional AI agent team management system based on the Otter Camp methodology. It enables you to coordinate multiple OpenClaw agents working together on complex projects with proper task assignment, progress tracking, quality control, and performance evaluation.

When to Use This Skill

Use this skill when you need to:

• Manage multiple AI agents working on the same project
• Implement structured workflows for complex tasks
• Track progress and ensure quality across agent teams
• Evaluate agent performance and optimize workflows
• Scale AI operations beyond single-agent capabilities

Core Features

Task Assignment & Coordination

• Intelligent task decomposition and assignment
• Agent role definition and specialization
• Dependency management between tasks
• Resource allocation and load balancing

Progress Tracking & Monitoring

• Real-time progress dashboards
• Milestone tracking and deadline management
• Automated status reporting
• Issue detection and escalation

Quality Control & Review

• Multi-layer quality assurance processes
• Peer review between agents
• Human-in-the-loop checkpoints
• Automated testing and validation

Performance Evaluation

• Agent performance metrics and scoring
• Workflow optimization recommendations
• Cost-benefit analysis of agent configurations
• Continuous improvement through learning

Usage Examples

Basic Team Setup

const teamManager = new AIAgentTeamManager({
  workspace: '/path/to/workspace',
  agents: ['xiaolv', 'laogou', 'xiaoqiu', 'xiaozhu'],
  methodology: 'otter-camp'
});

Task Coordination

await teamManager.assignTask({
  taskId: 'email-analysis-2026',
  description: 'Analyze 3,418 emails from QQ mailbox',
  assignee: 'xiaolv',
  reviewers: ['laogou'],
  deadline: '2026-03-10',
  qualityChecks: ['accuracy', 'completeness', 'formatting']
});

Performance Reporting

const report = await teamManager.generatePerformanceReport({
  period: 'last-30-days',
  metrics: ['tasksCompleted', 'qualityScore', 'efficiency', 'cost']
});

Integration Points

• Works seamlessly with existing OpenClaw agents
• Integrates with Git for version control
• Supports custom agent roles and specializations
• Compatible with all OpenClaw skill types

Best Practices

• Start with small teams (2-4 agents) and scale gradually
• Implement regular quality reviews and checkpoints
• Use human oversight for critical decisions
• Continuously optimize based on performance data
• Maintain clear documentation of team workflows

Security & Compliance

• All data remains in your local workspace
• No external API calls without explicit permission
• Full audit trail of all agent activities
• Compliant with enterprise security requirements