Purpose
This repository is the official integration and installable doctrine collection for AI Design Blueprint, with the goal of connecting the blueprint's principles, examples, and guides into various AI tools and agent runtimes, covering MCP, IDE rules, prompt files, and agent runtimes.
Repository contents
- shared/: specification files shared across tools
- mcp/: public MCP configuration and usage instructions
- docs/setup/: copy-first setup guides organized by tool
- cursor/, windsurf/, github-copilot/, gemini/: provider-specific instruction files
- open-weights/: static prompt packages for open-weight and local model workflows
- exports/: structured doctrine exports
Public endpoints
- Site:https://aidesignblueprint.com
- MCP：https://aidesignblueprint.com/mcp
- Developer docs:https://aidesignblueprint.com/en/for-agents
How to use
- Choose the corresponding setup guide in docs/setup/.
- Add the relevant files or MCP configuration to your own repository or client.
- If using MCP, initialize against https://aidesignblueprint.com/mcp .
- Run the first verification call: clusters.list()
- Run the second verification call: examples.search(query="orchestration visibility steering", limit=3)
Public MCP tools
Public retrieval tools (anonymous, read-only allowed):
- principles.list(cluster?)
- principles.get(slug)
- clusters.get(slug)
- examples.get(slug)
- principles.search(query, limit?)
- examples.search(query, principle_ids?, difficulty?, library?, limit?)
- guides.get(slug)
- guides.search(query, limit?)
Public signal tools (anonymous, opt-in writes required):
- signals.report(event_type, surface_used?, brief_context?, perceived_value?, workflow_stage?, would_recommend?, team_size?): records a value moment; only provide it after the user has explicitly expressed that something was useful, never call it automatically or silently.
- signals.feedback(task_type?, surface?, rating_clarity?, rating_usefulness?, what_helped?, what_missing?, would_use_again?, contact_email?, permission_to_follow_up?): explicit qualitative feedback; only call it when the user explicitly asks to leave feedback.
The signal tools only write the structured fields passed in and do not store prompts, code, or file contents. See the privacy policy for full data handling details.
Protected tools (require authentication, not part of the anonymous setup path):
- me.learning_path()
- me.coaching_context()
- architect.validate(implementation_context, ..., private_session?): Pro/Teams; scores agent code against 10 principles; setting private_session=true skips stored runs for that call.
- design.validate(implementation_context, ..., private_session?): Pro/Teams; interface mirror, scores the rendered frontend output against 8 experience design laws (separate weekly bucket).
- spec.validate(implementation_context, ..., private_session?): Pro/Teams; build content perspective, scores written specs against 8 spec quality laws (separate weekly bucket).
- team.summarize(days_back?, private_session?): Pro/Teams; usage reflection across the three validator perspectives and recommended follow-up assets.
- me.add_evidence(course_slug, stage_id, note)
Feedback and value signal rules
- Only call signals.report after the user explicitly states that some content was useful. Never call it automatically or silently. After an explicit success signal, provide it at most once per session.
- Only call signals.feedback when the user explicitly asks to leave feedback. Never prompt proactively.
- Never include proprietary code, file contents, or confidential information in brief_context.
Governance badges
Used to indicate that an agent or repository follows the Blueprint specification.
- Free badge: paste into README.md (no account required).
- Pro badge: run architect.validate() via MCP. The response includes run_id, badge_url, and review_url. The Pro badge shows the tier (Governed · X/Y or Reviewed · X/Y) and links to a public-readiness review page. Requires a Pro or Beta account.
What is intentionally not provided at this time
- No public OpenAPI schema
- Except for MCP and static assets,no public HTTP API contract
- No CLI installer
- No speculative partner-exclusive distribution
Source of truth
This repository is intended to mirror the published specification public contract on aidesignblueprint.com. Before publishing changes here, confirm that /agent-assets/[slug] matches the files committed in this repository.
