Skip to content

Port content improvements from Claude Code authstack#11

Closed
saif-at-scalekit wants to merge 1 commit into
mainfrom
cross-pollinate/cursor-content-improvements
Closed

Port content improvements from Claude Code authstack#11
saif-at-scalekit wants to merge 1 commit into
mainfrom
cross-pollinate/cursor-content-improvements

Conversation

@saif-at-scalekit
Copy link
Copy Markdown
Collaborator

What

Ports content patterns from the Claude Code authstack to Cursor, adapting them to Cursor's native .mdc rule format.

Changes

New .mdc rules (5 files)

Rule Plugin Purpose
terminology.mdc agent-auth Enforces connector/connection/connected-account/tool terminology
live-metadata-first.mdc agent-auth Treats live API metadata as source of truth over static docs
tool-selection.mdc agent-auth Keeps tool sets minimal for better LLM accuracy
redirect-urls.mdc full-stack-auth Catches redirect URL mismatches before auth failures
terminology.mdc full-stack-auth Enforces consistent product naming across auth surfaces

Skill improvements (8 files across all 5 plugins)

  • Mental model sections added to agent-auth and full-stack-auth core skills
  • When to switch skills routing tables added to all 5 plugin entry skills
  • Deep reference backlinks added to building-agent-mcp-server and production-readiness skills
  • Decision table + critical prereqs added to mcp-auth entry skill
  • Stale beta SDK version pin removed from agent-auth skill and setup-scalekit agent
  • Broken CONNECTORS.md link fixed in agent-auth skill
  • Stale Reddit link removed from modular-sso skill

Why

The Claude Code authstack developed content patterns (terminology rules, skill routing, mental models, live-metadata-first) that improve agent accuracy and reduce context confusion. This PR brings those patterns to Cursor while using Cursor-native .mdc format for rules.

Related

New .mdc rules:
- agent-auth: terminology, live-metadata-first, tool-selection
- full-stack-auth: redirect-urls, terminology

Skill improvements across all 5 plugins:
- Add mental model sections (agent-auth, full-stack-auth)
- Add 'When to switch skills' routing tables (all plugins)
- Add deep reference backlinks (building-agent-mcp-server, production-readiness)
- Add decision table and critical prereqs (mcp-auth)
- Remove stale beta SDK version pin (agent-auth, setup-scalekit)
- Fix broken CONNECTORS.md link (agent-auth)
- Fix stale Reddit link (modular-sso)
@saif-at-scalekit
Copy link
Copy Markdown
Collaborator Author

Superseded by #12 which includes these content improvements plus the 5→2 AgentKit/SaaSKit consolidation.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants