lightspeed-ai-readiness-orchestrator

Files (8)

• SKILL.md
• openai.yaml · agents/
• icon.svg · assets/
• client-project-wizard.md · references/
• lightspeed-service-positioning.md · references/
• output-standards.md · references/
• red-flags.md · references/
• skill-routing-rules.md · references/

SKILL.md (first lines)

# LightSpeed AI Readiness Orchestrator

## Purpose

Coordinate LightSpeed AI readiness client projects by routing the work through the installed specialist skills:

1. `ai-readiness-assessor`
2. `ai-governance-documentor`
3. `content-collection-planner`
4. `ai-chatbot-planner`
5. `lightspeed-ai-readiness-router`

This skill is the front door. It must orchestrate and sequence the specialist workflows rather than duplicate them.

## Default workflow

Always start with the router logic unless the user explicitly asks for one specialist workflow.

1. Create a source material inventory.
2. Confirm project context.
3. Route the project to the correct specialist workflow.
4. Enforce the normal sequence:
   - readiness assessment
   - governance discovery and documentation
   - content collection
   - chatbot planning
   - roadmap or proposal pack
5. Produce Markdown files or Markdown-ready sections by default.
6. Create a combined project pack when the user asks for files or a ZIP.

## Intake rules

Before routing, collect the minimum useful context:

- client name
- website URL
- sector or industry
- project type
- project stage: new lead, existing client, retainer client or internal LightSpeed initiative
- available source files or links
- desired output
- budget or timeline constraints, if known
- Figma design system URL, if available
- GitHub repositories, if relevant
- other useful URLs such as analytics dashboards, Search Console, staging site, sitemap, brand files or content inventory

Ask for uploaded files before each major workflow. Be specific about what is needed.

## Source material inventory

When source files are available, list them before asking detailed questions.

Classify each source as:

- approved source of truth
- useful but needs review
- draft or incomplete
- outdated or risky
- suitable for chatbot grounding
- not suitable for chatbot grounding
- legal or privacy review recommended

Do not invent evidence. If evidence is missing, ask focused follow-up questions.

## Routing logic

Use `references/skill-routing-rules.md` when choosing the workflow.

Use this default routing:

- If the user asks whether a website is ready for AI, route to `ai-readiness-assessor`.
- If the user asks about governance, policy, AI use rules, risk, approvals or responsible AI operations, route to `ai-governance-documentor`.
- If the user needs to collect content, organise source material, create a content request or identify content gaps, route to `content-collection-planner`.
- If the user asks about chatbot scope, AI Engine, source lists, escalation, fallback wording, launch gates or chatbot system prompts, route to `ai-chatbot-planner`.
- If the user asks broadly to start a client AI readiness project, use this orchestrator and the router sequence.

## Strict mode

Switch to strict interactive mode when the client operates in any high-risk or high-trust context, including:

- healthcare
- finance
- legal
- children or vulnerable users
- education
- insurance
- employment
- public sector
- regulated ecommerce
- sensitive personal data
- regulated advice or high-impact decisions

In strict mode:

1. Pause broad recommendations.
2. Ask additional risk, privacy, data handling, ownership and approval questions.
3. Do not recommend chatbot implementation until source content, escalation and governance are adequate.
4. Include a legal/privacy disclaimer in client-facing governance and chatbot outputs.
5. Treat missing privacy, cookie or terms documents as a warning.

## Output standards

Use `references/output-standards.md` for default document structure.

Default output should be Markdown suitable for Google Docs, GitHub or later PDF/DOCX conversion.

Separate:

- client-facing content
- internal LightSpeed notes
- open questions
- risks and assumptions
- recommended next step

Every major output should start with:

- value
- risks
- next step

From lightspeedwp/.github, GPL-3.0.

More from lightspeedwp/.github