Use when studying how superpowers skills are structured to understand why each section exists
How this skill is triggered — by the user, by Claude, or both
Slash command
/skill-factory-toolkit:sft-skill-anatomyThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Structural breakdown of three representative superpowers skills. Maps every section to the writing-skills authoring guide and explains the design rationale behind each structural choice.
Structural breakdown of three representative superpowers skills. Maps every section to the writing-skills authoring guide and explains the design rationale behind each structural choice.
Look up the skill type closest to what you are building, then consult the summary table below. For full section maps, removal-impact analysis, and supporting file breakdowns, read skill-analyses.md.
| Skill | Type | Word Count | Supporting Files | Key Insight |
|---|---|---|---|---|
| brainstorming | Technique | ~1600w | 7 (visual companion, scripts, reviewer prompt) | Layered defense: HARD-GATE + anti-pattern + checklist + flow + user gate |
| test-driven-development | Discipline | ~1100w | 1 (testing-anti-patterns.md) | Anti-rationalization architecture: Iron Law + verify steps + comparisons + table + flags |
| writing-skills | Meta | ~1800w | 6 (best practices, persuasion, testing, graphviz, renderer, example) | Self-exemplification: follows every convention it teaches |
| Dimension | brainstorming | test-driven-development | writing-skills |
|---|---|---|---|
| Type | Technique | Discipline-Enforcing | Meta |
| Enforcement | Flexible within steps, rigid step order | Rigid everywhere | Rigid for rules, flexible for application |
| Has Iron Law | No (uses HARD-GATE instead) | Yes | Yes |
| Has rationalization table | No (single anti-pattern block) | Yes (11 rows) | Yes (8 rows) |
| Has Red Flags list | No | Yes (13 items) | No (uses STOP section) |
| Has flowchart | Yes (process flow) | Yes (cycle) | Yes (decision) |
| Has verification checklist | No (uses user review gate) | Yes (8 items) | Yes (full TDD checklist) |
| Defensive layers | 5 | 5 | 5 |
| Primary risk | Premature implementation | Rationalization to skip TDD | Undiscoverable or untested skills |
Patterns that hold across all three skills:
| Principle | Evidence |
|---|---|
| Defense in depth | Every skill has 5+ independent enforcement mechanisms |
| Front-load critical content | Hard gates and iron laws appear in the first 20% of each skill |
| Close specific loopholes, not general ones | Rationalization tables address verbatim excuses from baseline testing |
| Supporting files for heavy reference only | SKILL.md stays focused; large content loads on demand |
| Flowcharts only for non-obvious decisions | All three use flowcharts for state machines or decision trees, never for linear steps |
| Self-exemplification in meta skills | writing-skills follows every convention it teaches |
| Token budget awareness | Discipline skills stay minimal; technique skills scale sections to complexity |
| File | Purpose |
|---|---|
| skill-analyses.md | Full section maps, removal-impact analysis, and supporting file breakdowns for all three skills |
npx claudepluginhub steltz/skill-factory --plugin skill-factory-toolkitGuides designing tightly focused, composable skills as mental models, principles, techniques, and processes at various abstraction levels for dynamic integration like a knowledge graph.
Creates and validates skills using test-driven development with pressure scenarios and subagents. Use when authoring new skills, editing existing ones, or verifying skill behavior.
Uses test-driven development to create, edit, and verify skills for agent behavior. Write pressure scenarios, watch agents fail, then document rules to fix behavior.