Skip to content

rampstackco/claude-skills / skills/skill-creation-walkthrough

skill-creation-walkthrough

Verified

Step-by-step guide for creating your own Claude Skills, from deciding whether a skill is the right tool to writing the SKILL.md file, structuring reference material, and making it trigger reliably. Use when you want to package a workflow, framework, or repeated task into a reusable Skill, when an existing skill is not triggering or not loading the right context, when you are auditing a skill that is underperforming, or when you want to publish a skill for others. Also triggers when someone asks "how do I make a skill" or "what makes a good skill". Useful for individuals, teams, and anyone publishing skills publicly.

FileKindSize
README.mdmarkdown1.8 KB
SKILL.mdskill md13.6 KB
references/description-cookbook.mdmarkdown9.2 KB
references/methodology-vs-implementation.mdmarkdown7.6 KB
references/skill-template.mdmarkdown5.6 KB
# skill-creation-walkthrough

The skill is [SKILL.md](SKILL.md). Generated by `scripts/build-tested-on.mjs` from openaddict.com. CI checks it against the site, so edit the script, not this file.

## Tested on OpenAddict

Measured on [OpenAddict](https://openaddict.com/skills/rampstackco-skill-creation-walkthrough): the same skill authoring tasks, with this skill loaded and without it, on each model below.

- **Tier:** Tier 1, the first tasks built for this job.
- **Date:** pinned at commit `0479242` on 2026-08-30; newest run 2026-09-26.
- **Full results:** [openaddict.com/skills/rampstackco-skill-creation-walkthrough](https://openaddict.com/skills/rampstackco-skill-creation-walkthrough)

| Model | Via API | In Claude Code |
|---|---|---|
| Claude Haiku 4.5 | Holds | Holds |
| Gemini 3.1 Flash Lite | Mixed, 2 results: No measured effect 1, Holds 1 | not run |
| GPT-5 mini | Holds | not run |
| Claude Fable 5 | not run | Mixed, 3 results: No measured effect 2, Scored worse 1 |
| Claude Opus 5 | not run | Could not measure |
| Claude Sonnet 5 | not run | No measured effect |
| Claude Fable 5.1 | not run | Mixed, 2 results: Could not measure 1, No measured effect 1 |
| Claude Opus 5.5 | not run | Could not measure |

Each word is the verdict the site prints. Holds: the skill helped, by more than the margin the site needs before it calls a result. No measured effect: no difference large and certain enough to call. Could not measure: the test could not show a difference either way. Scored worse: the model did worse with the skill loaded. Mixed: the model was measured more than once and the results differ, so each is counted. Not run: this model was not run this way.

OpenAddict is run by the people who publish this catalog, and these results are theirs. Nobody outside checked them.
rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit.Stars: 471kUpdated Claude CodeCodexVerifiedInstall
rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure.Stars: 471kUpdated Claude CodeCodexVerifiedInstall
rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics.Stars: 471kUpdated Claude CodeCodexVerifiedInstall
ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first.Stars: 194kUpdated Claude CodeCodexVerifiedInstall