Mission Control Skill Ecosystem Builder

Mission Control Skill Ecosystem Builder

Purpose

Route skill design and packaging work through Mission Control so reusable instructions stay testable, discoverable, and host-safe.

The Codex chat agent is not the Mission Control Manager. It is the bridge between the user and the Mission Control Manager.

Use when

• The user wants a new skill or skill pack.
• Existing skills need cleanup, metadata, tests, or marketplace packaging.
• A skill should work across Codex, Claude Code, or other agent hosts.

Workflow

1. Ask Mission Control to identify the skill goal, trigger conditions, required assets, and host compatibility.
2. Request a skill contract with frontmatter, workflow, examples, failure modes, and safety boundaries.
3. Validate the skill with package/catalog checks.
4. Summarize changed assets and any host reload requirements.

Mission Control calls

Tools:

• mission_control_start_task
• mission_control_get_status

Resources:

• mission-control://projects/{project_id}/status
• mission-control://projects/{project_id}/agent-contracts
• mission-control://projects/{project_id}/validation-summary

User-facing output

• Explain the skill purpose, triggers, files changed, validation run, and how the host discovers it.
• Keep examples concrete and copy-pasteable.

Approval behavior

Ask before publishing, installing globally, or overwriting user-owned skill assets.

Never do

• Do not copy third-party skill text wholesale.
• Do not ship a skill without clear trigger criteria.
• Do not claim host compatibility without validation.

Failure and fallback

If package validation is unavailable, produce a draft skill and mark host compatibility as unverified.

Example invocation

Use Mission Control to create a reusable skill for release readiness checks.