safe-skill-creator

Skill design, testing, iteration for Claude. Build new skills from scratch, refactor existing skills, optimize triggering accuracy, evaluate quality.

Grounded in four design strategies: Processing · Mediation · Forgetting · Integrity.

Set of Files

File	Purpose
SKILL.md	Router — load order
safe-skill-creator.md	Content — philosophy, 8-phase lifecycle, integrity checks
safe-skill-creator.skill	Packaged archive for claude.ai upload
references/claude-env.md	Claude environment specific tests and optimizations
.github/workflows/build-skill.yml	CI workflow — auto-builds and publishes safe-skill-creator.skill on push
README.md	Explanatory instructions and overview for this software package

Install

• claude.ai web platform: upload safe-skill-creator.skill via skill settings (recommended). Or add safe-skill-creator-SKILL.md contents to Personal Preferences via Settings > General.

• Claude Code desktop app: copy contents of this folder into ~/.claude/skills/safe-skill-creator/.

• Other platforms: adapt this set of files to your harness+model. Star or Fork the git repo if you like.

8-Phase Lifecycle

1. Capture intent
2. Interview + research
3. Write SKILL.md
4. Write test cases
5. Run + evaluate
6. Improve
7. Optimize description triggering
8. Package + present

When It Triggers

• "turn this into a skill", "make a skill for X"
• "why isn't my skill firing"
• Skill refactor, benchmark, description tuning

Undertriggers by default. If in doubt, use this skill as meta-framework for skill creation and testing.

Peer Skills

• prompteng — parent skill; §6 persistence formats
• captureng — capture sessions into skills
• packageng — package finished skills into .skill

Companion Files

Available in - ecological-codes/user-prefs

• agent.md

• trusted-hosts.md

Style Convention

Directives use the following section headers with numbered lists, shared across peer skills:

• [RULES] — enforceable constraints applied at runtime.
• [ACTIONS] — autonomous steps agent executes in normal workflow.
• [HUMAN ACTIONS] — UI actions; agent skips, cannot delegate.

License

See LICENSE. (C) Copyright 2026 - Sameer Khan - Various and Several Rights Reserved.

---

README.md v2.2.0 - Human Approved