Your habits
Store personal skills under ~/.claude/skills/. Optimize for your own workflow without promising compatibility to others.
Community-built tools that extend Claude Code. Every tool verified against its public repository.
A public catalog helps you discover procedures and patterns. An install count, leaderboard position, or marketplace rating does not prove that a skill fits your client, repository, risk level, or maintenance model.
Store personal skills under ~/.claude/skills/. Optimize for your own workflow without promising compatibility to others.
Commit project skills under .claude/skills/ with a named owner, review path, assumptions, and retirement signal.
Publish through a maintained repository or plugin when users need compatibility and release support for a specific tool.
Treat the catalog as a source of candidates. Consume, fork, specialize, extract a pattern, or reject after local evaluation.
Review the selected revision, SKILL.md, supporting files, scripts, dependencies, and tool grants. Claude Code-specific fields may behave differently in another client.
npx skills add <owner/repo> Security: treat a downloaded skill like executable code. Ownership: treat it like situated context whose assumptions and maintainer must be recorded.
Skill ownership guide →No tools match your filters.
| Persona | Recommended Tools | Rationale |
|---|---|---|
| Solo developer | ccusage + claude-code-viewer | Cost awareness + session history review |
| Small team (2-5) | ccusage + Conductor or multiclaude | Cost tracking + parallel development |
| Enterprise | ccusage (MCP) + Entire CLI | Programmatic cost data + audit trails |
| Python-centric | ccburn + Claude Chic | Native Python ecosystem tools |
| Multi-agent user | Toad or Conductor | Unified agent management |
| Config-heavy setup | claude-code-config + AIBlueprint | TUI config management + scaffolding |
Not sure whether to use MCP servers or CLI tools for a given integration? The decision guide covers four key questions, a 15-row guidance table by situation, and token cost analysis.
MCP vs CLI: Which to use? →As of March 2026, community tooling has notable missing pieces:
No GUI for creating or editing .claude/skills/. Must edit YAML/Markdown manually.
No GUI for managing hooks in settings.json. Requires JSON editing.
No single dashboard combining config, sessions, cost, and MCP management.
No way to measure token cost attributable to each MCP server individually.
No tool syncs Claude Code config across machines. Manual copy of ~/.claude/ required.
Entire CLI (launched Feb 2026) provides rewindable checkpoints with full context replay.