Skills child-agent-ops
🤖

child-agent-ops

Safe 🌐 Network access⚙️ External commands

Manage child agents with standardized workflows

Multi-agent projects often face coordination chaos. This skill provides clear role definitions, checklists, and templates so leader, implementer, and reviewer agents can collaborate efficiently without confusion.

Supports: Claude Codex Code(CC)
📊 70 Adequate
1

Download the skill ZIP

2

Upload in Claude

Go to Settings → Capabilities → Skills → Upload skill

3

Toggle on and start using

Test it

Using "child-agent-ops". Launch a codex child agent to implement user authentication

Expected outcome:

  • [Child Agent Launch]
  • - Type: Implementation
  • - Task ID: AUTH-001
  • - Scope: User login/logout, session management, password reset
  • - Completion: User authentication works with test coverage >80%
  • - Estimated time: 2 hours
  • - Report checkpoint: Every 30 minutes

Using "child-agent-ops". Report status for the database migration task

Expected outcome:

  • [Task Status Report]
  • - Task ID: DB-042
  • - Current: Schema migration complete, data validation in progress
  • - Next: Rollback script review, production deployment
  • - Issues: Minor performance degradation on staging (investigating)

Using "child-agent-ops". Request review for security configuration changes

Expected outcome:

  • [Review Request]
  • - Risk Level: HIGH
  • - Changes: firewall rules, IAM policy updates
  • - High-risk operations: Production access policy modifications
  • - Attached: Security impact analysis, test results

Security Audit

Safe
v5 • 1/16/2026

This is a pure documentation skill containing only markdown guidelines for child agent operations. All 20 static findings are FALSE POSITIVES: hash values were misidentified as cryptographic weaknesses, Japanese text was flagged as crypto algorithms, and documentation templates with code examples triggered shell execution warnings. No executable code, scripts, network operations, filesystem access, or command execution exists. The skill provides only templates and workflow guidance.

2
Files scanned
262
Lines analyzed
2
findings
5
Total audits
Audited by: claude View Audit History →

Quality Score

38
Architecture
100
Maintainability
87
Content
30
Community
100
Security
83
Spec Compliance

What You Can Build

Coordinate multi-agent teams

Break down complex projects into parallel tasks, assign specialized agents, and monitor progress through standardized checkpoints and reporting.

Execute delegated tasks

Follow defined scope and completion criteria, report risks immediately, and coordinate with reviewers for code changes.

Ensure safe operations

Enforce pre-approval for high-risk operations like security changes, data deletion, or billing modifications.

Try These Prompts

Launch Child Agent
Launch a codex child agent for [TASK-NAME]. Purpose: [one-line goal]. Completion criteria: [specific measurable outcome]. Estimated time: [duration]. Scope: [boundaries]. Report progress every checkpoint.
Report Task Status
Report current status for [TASK-NAME]. Current state: [what is done]. Next: [what remains]. Issues: [blockers or risks]. Attach codex launch logs and execution logs.
Request Review
Request review from a reviewer agent. Changes: [files modified]. Risk level: [low/medium/high]. High-risk operations: [none or describe]. Attach implementation summary and testing notes.
Handle Blocker
BLOCKER: Cannot proceed with [TASK-NAME] due to [reason]. Options considered: [what was tried]. Recommendation: [suggested path forward]. Awiting leader decision before continuing.

Best Practices

  • Always document purpose and completion criteria before launching child agents
  • Report risks and blockers immediately rather than waiting for completion
  • Use codex launch logs with task IDs to track agent activities
  • Require reviewer agent approval for any code changes before merging

Avoid

  • Launching child agents without clear completion criteria
  • Waiting until task completion to report blockers or issues
  • Skipping reviewer checkpoints for security-sensitive changes
  • Allowing parallel updates without leader coordination

Frequently Asked Questions

Which AI tools are compatible with this skill?
This skill works with Claude, Codex, and Claude Code. The codex tool is required for launching child agents.
What are the limits on child agent parallelism?
No hard limits, but coordination costs increase with more agents. Leader should balance parallelism against communication overhead.
How does this integrate with existing project workflows?
Add the skill to your agent context. Use the provided templates for all child agent launches and status reports.
Is data from child agent sessions stored securely?
This skill only provides templates. Storage and security depend on your environment. Attach logs to your existing documentation system.
Why is my child agent task not completing?
Check completion criteria clarity. Ensure scope is bounded. Verify the agent has all necessary permissions and tools available.
How is this different from standard agent prompting?
This skill enforces structured workflows with roles, checkpoints, and templates. It prevents coordination gaps in multi-agent projects.

Developer Details

File structure

📄 SKILL.md