← ClaudeAtlas

claude-config-self-development-researchlisted

Research whether Simon's Claude config, his sk/sk-work skills, the third-party packs and the tools they depend on have fallen behind current community-validated practice — then propose the worthwhile changes for approval. Audits ~/.claude against primary sources and high-adoption repos, checks installed packs and CLI deps for drift and deprecation, and separately surfaces proven workflow practices he is not using yet. Use for "what am I missing", "is my Claude setup current", "research better ways to do this", "audit my config against best practice", or on a quarterly refresh. Proposes only; never applies a change without his yes.
skowalskidev/dotclaude · ★ 2 · AI & Automation · score 81
Install: claude install-skill skowalskidev/dotclaude
# Self-development research Simon's config is a living system. It drifts in three directions at once, and none of them announce themselves: - **The platform moves.** Claude Code ships features that make a hand-built mechanism redundant, or quietly changes one his config depends on. - **The packs move.** Third-party skills are cloned once and then sit there. Upstream keeps shipping. - **The config rots in place.** A rule claims a behaviour that no longer happens. A doc names a mechanism that was replaced. Nothing fails, so nothing surfaces it. This skill goes and looks. It produces a proposal, not a change. ## The bar — the whole point of this skill Simon's instruction: **not so loose that it suggests minute half-improvements, not so tight that it misses genuinely new developments.** The bar is what makes this skill useful rather than noise, so apply it hard. **A finding qualifies only if at least one is true:** 1. Something is **silently broken or not doing what it claims** — a rule that never loads, a guard that cannot fire, a doc that describes a mechanism that no longer exists. 2. The platform now has a **first-class feature** that replaces something hand-built, or that closes a gap his own rules say is open. 3. A dependency, pack or tool is **stale enough to matter** — months behind, deprecated, abandoned, or superseded by something with clearly wider adoption. 4. A **new capability** exists that would change how he works, not just how something is s