Byte Sizesandy.exe — learning
← Back to worlds

Claude Code / Skills

Claude Code features, custom skills, slash commands, CLAUDE.md conventions.

The collection74 saves · newest SEP 05
Player 1 · this device
0 / 74 adopted
  1. LV.1A PRD Builder Slash Command for Claude Code (TDD + Spec-Driven + Recursive Testing)
  2. LV.2Refactor a Bloated Controller With Claude Code Plan Mode and Mid-Run Steering
  3. LV.3A Self-Built Remote Control for a Coding Agent
  4. LV.4Use Claude Skills to Stand Out at Work (Team Deployment + Career Strategy)
  5. LV.5Keep CLAUDE.md Minimal, Split Everything Else Into a .claude/ Folder
  6. LV.6The Four .claude/ Files That Matter: CLAUDE.md, Rules, settings.json, Hooks
  7. LV.7Claude Code Status Line: Folder, Branch, Context Bar, Model at a Glance
  8. LV.8Four Add-Ons Recommended for a Coding Agent
  9. LV.9Give the Agent a Prebuilt Map of the Codebase
  10. LV.103 Claude Code Features Most People Miss: Sub-Agents, /status, and Auto-Skill Hooks
  11. LV.114 MCP Tools That Make Claude Code 10x Smarter (Context7, Sequential Thinking, Superpowers, Plan Mode)
  12. LV.12The Repo Layout for Claude: Three Root Files Plus a Config Folder
  13. LV.13Give a Coding Assistant Eyes by Delegating to a Multimodal Model
  14. LV.14Eight Claude Code Skills, Listed by Category
  15. LV.15Ask a Side Question Without Interrupting the Main Task
  16. LV.16What Is Actually Inside a Skill Folder
  17. LV.17Give the Agent a Completion Condition Instead of Supervision
  18. LV.18Write a Handoff File Instead of Compacting
  19. LV.19An Engineer's Open-Sourced Skills Repository
  20. LV.206 Claude Code Plugins Almost Nobody Is Using
  21. LV.21Ask for a Mental Model Instead of Telling AI to Fix It
  22. LV.22"Lock It In": Claude Skill That Properly Captures and Packages Your Workflows
  23. LV.23Keep a Running Decision Log While Implementing a Spec
  24. LV.24Five Coding Agent Commands Most People Miss
  25. LV.25Promote Repeated Lessons Into Permanent Memory
  26. LV.26A Coding Agent Workflow Built on Plan, Delegate, and Prove
  27. LV.27One Orchestrator Skill That Routes to Specialists
  28. LV.28How the Creator of Claude Code Uses Scheduled Loops
  29. LV.29End a Correction With Write the Rule Down
  30. LV.30Build an Agent From a Requirements Doc and an Instructions File
  31. LV.31Keep Context Under Forty Percent to Avoid Degradation
  32. LV.32The Project Instructions File as the Project's System Prompt
  33. LV.33Fewer Instructions in the Project File, Not More
  34. LV.34CLAUDE.md + handoff.md + Start/Stop Hooks: Claude Code Always Knows What to Work On Next
  35. LV.35Five Community-Built Claude Code Plugins, Ranked
  36. LV.36Raise Reasoning Effort Only for Genuinely Hard Tasks
  37. LV.37Pass Arguments and Set Invocation Rules on a Skill
  38. LV.38Audit Stale Memories, Skills, and Configs With a Prune Routine
  39. LV.39Clear the Session After Every Finished Task
  40. LV.40Use the Strongest Model as an Orchestrator, Not a Worker
  41. LV.41Two Skills That Create and Retire Your Other Skills
  42. LV.42Route Tasks to the Cheapest Model That Can Do Them
  43. LV.43A Spec, Build, Review Loop Made of Three Skills
  44. LV.44Use a Canary Instruction to Detect Context Degradation
  45. LV.453 Levels of Claude Code Mastery: Extension → CLI → Multi-Agent Expert
  46. LV.463-Step Token Management: CLAUDE.md + Context Folder + Fresh Chat Per Task
  47. LV.47Claude Code Workflow: Explore → Plan → Code → Commit (Don't Code First)
  48. LV.48/visual-plan and /visual-recap: MDX Skills for Visual Planning + Code Review with Claude
  49. LV.49Make a Coding Agent Ask Whether the Code Should Exist
  50. LV.50Build a Personal Claude Harness in Under an Hour: One Folder, One CLAUDE.md
  51. LV.51Use a Symlink to Sync AI Context Files Across Tools (CLAUDE.md, agents.md)
  52. LV.52"Last 30 Days" Claude Skill: Restricts Research to Recent Sources Only
  53. LV.53Free Claude Skill: 300-Check Codebase Audit Across UI/UX, Performance, and Security
  54. LV.54Claude Code Artifacts: Auto-Generated Project Summaries to Share with Your Team
  55. LV.55A .claude Folder Built Like an Onboarded Engineer
  56. LV.56The 170k-Star CLAUDE.md File That Stops Claude Code From Over-Engineering
  57. LV.573 Claude Code Skills for Hard Problems: Watchdog, Planarbiter, and RTFD
  58. LV.58Claude Code's Hidden Fullscreen Rendering Mode
  59. LV.59Structure an Obsidian Vault So a Coding Agent Can Navigate It
  60. LV.60The "Ponytail" Claude Skill That Cuts Vibe-Coded Bloat by 90%
  61. LV.613 Claude Plugins That Fix the Disconnected AI Problem
  62. LV.62Reset Claude Code Context with handoff.md Instead of /compact
  63. LV.63Anthropic Internal Claude Code Masterclass: Key Takeaways
  64. LV.645 Claude Code Skills That Teach It How Obsidian Actually Works
  65. LV.65Convert PDFs to Markdown Before Feeding Them to an Agent
  66. LV.66Planner-Executor Loop: Top Model Plans, Cheaper Model Does Grunt Work
  67. LV.67Split a Bloated Instructions File Into Path-Scoped Rules
  68. LV.686 Ways You Are Wasting Tokens in Claude (and the Fixes)
  69. LV.695 Claude Code Skills Picked From Testing Over 100
  70. LV.70Claude Code's /checkup Command: 7 Fixes That Cut Context Bloat
  71. LV.71A Claude skill that heals your other skills after each session
  72. LV.72Self-Improving Agent Skills From an Open Source Optimizer
  73. LV.73Chaining Plan, Goal, Background Agents, and Review
  74. LV.74Six Questions Your ARCHITECTURE.md Should Answer