// Track

Skill Engineering Mastery

Author Phenomenal Skills That Remember and Evolve

The mastery sequel to Skills, Hooks & Slash Commands. Go past authoring basics to the discipline that turns a skill library into an authority moat: trigger contracts that fire proactively and refuse the adjacent wrong case, PreToolUse gates and PostToolUse verifiers that enforce quality the model cannot override, a persistent memory brain wired into your skills, the correction→rule self-evolution loop, and the sanitization gates that let you ship outward-facing skills without leaking. Ten hands-on lessons, each with a BuildChallenge you author yourself.

Recommended: Complete Skills, Hooks & Slash Commands first

10 lessons~87 min total

Lessons are shown in recommended order. Complete them in sequence for the best experience — or jump to any lesson.

Lesson 661·8 min read

The Authority Bar: What Makes a Skill Phenomenal

An adequate skill runs when you call it by name; a phenomenal skill fires itself at the right moment and refuses the wrong one — and the entire gap is authorship, not effort.

Lesson 662·9 min read

Anatomy of a Production Skill

A production skill is a directory, an entrypoint, two load-bearing frontmatter fields, a progressively-disclosed body, and optional supporting files — and knowing which part the model reads when is what keeps a rich skill cheap.

Lesson 663·9 min read

Triggers That Actually Fire

At decision time the model has read your skill's description and nothing else — so the proactive-invocation contract, and the rationalization counters that hold it back, both live or die in that one field.

Lesson 664·9 min read

PreToolUse Hooks: Gating Before the Action

A PreToolUse hook is the last checkpoint before a tool runs — where you allow, ask, or deny with a contract the model cannot reason its way past.

Lesson 665·9 min read

PostToolUse Hooks: Verification After the Action

A PostToolUse hook cannot stop what already happened — but by verifying the aftermath and feeding the result back as context, it turns a silent-broken edit into an immediate correction loop.

Lesson 666·9 min read

Installing a Memory Brain

A skill without memory starts from zero every run — give it a store it reads at the start and writes to at the end, and it stops repeating its own mistakes across sessions.

Lesson 667·8 min read

The Recall Loop: Making Skills Remember

A phenomenal skill reads memory before it acts — query the store, rank by relevance, inject only the top few — so it never relearns the same lesson twice.

Lesson 668·8 min read

Self-Evolution: The Correction-to-Rule Loop

A correction is the skill's training signal — capture it, categorize it, write it to lessons.md and memory, and escalate the ones that keep recurring.

Lesson 669·9 min read

Sanitization Gates: Shipping Skills Safely

Outward-facing skills must not leak — a generation-time gate blocks credentials, PII, infra, and internal names before commit, with CI as the backstop.

Lesson 670·9 min read

The Compounding Skill Library

A curated library of phenomenal skills is an authority moat — each skill makes the next cheaper to write, and pruning is what keeps the flywheel spinning forward.