{"slug":"gentle-ai-skill-improver-2","title":"gentle-ai-skill-improver","summary":"Trigger: improve skills, audit skills, refactor skills, skill quality. Audit and upgrade existing LLM-first skills.","platform":"Claude","tags":[],"authorName":"LLM Mart","authorSlug":"llm-mart","score":0,"source":"github","price":null,"verified":false,"createdAt":"2026-09-20T08:00:32.903692Z","repo":{"url":"https://github.com/Gentleman-Programming/gentle-shell","stars":1123,"forks":158,"license":"MIT","updatedAt":"2026-09-27T20:40:53Z"},"bodyHtml":"<hr>\n<h2>name: gentle-ai-skill-improver\ndescription: \"Trigger: improve skills, audit skills, refactor skills, skill quality. Audit and upgrade existing LLM-first skills.\"\nlicense: Apache-2.0\nmetadata:\nauthor: gentleman-programming\nversion: \"1.0\"</h2>\n<h2>Activation Contract</h2>\n<p>Use this skill when auditing, refactoring, normalizing, or improving existing <code>SKILL.md</code> files. Use <code>gentle-ai-skill-creator</code> when creating a brand-new skill from a reusable pattern.</p>\n<h2>Hard Rules</h2>\n<ul>\n<li>Read <code>docs/skill-style-guide.md</code> first and treat it as the normative style contract.</li>\n<li>Treat <code>SKILL.md</code> as the source of truth; preserve author intent, critical rules, activation semantics, and output requirements.</li>\n<li>Use <code>.atl/skill-registry.md</code> as an index of skill names, triggers, scopes, and exact paths when available.</li>\n<li>Default to audit-only. Modify files only when the user explicitly asks to apply improvements.</li>\n<li>Never delete meaningful content silently; move long explanation, examples, templates, or schemas into local <code>references/</code> or <code>assets/</code>.</li>\n<li>Do not invent triggers, policies, or domain rules. Mark ambiguous cases for human review.</li>\n</ul>\n<h2>Decision Gates</h2>\n<table>\n<thead>\n<tr>\n<th>Situation</th>\n<th>Action</th>\n</tr>\n</thead>\n<tbody>\n<tr>\n<td>Missing or invalid frontmatter</td>\n<td>Fix <code>name</code>, quoted one-line <code>description</code>, <code>license</code>, and <code>metadata</code></td>\n</tr>\n<tr>\n<td>Skill reads like tutorial docs</td>\n<td>Convert to runtime instructions and move background to <code>references/</code></td>\n</tr>\n<tr>\n<td>Body exceeds budget</td>\n<td>Preserve rules, move examples/background to supporting files</td>\n</tr>\n<tr>\n<td>Branching logic hidden in prose</td>\n<td>Convert to a compact decision table</td>\n</tr>\n<tr>\n<td>Rules conflict or intent is unclear</td>\n<td>Report the issue; do not rewrite that rule automatically</td>\n</tr>\n</tbody>\n</table>\n<h2>Execution Steps</h2>\n<ol>\n<li>Read <code>docs/skill-style-guide.md</code>.</li>\n<li>Read <code>.atl/skill-registry.md</code>; use listed paths to select skills. If missing, scan known skill directories for <code>*/SKILL.md</code>.</li>\n<li>For each selected skill, audit metadata, trigger clarity, section order, body budget, actionability, decision gates, output contract, and local references.</li>\n<li>Return an audit report grouped by skill with severity and exact proposed changes.</li>\n<li>In apply mode, edit only safe issues, preserve content, create supporting files when needed, then refresh or request <code>/skill-registry:refresh</code>.</li>\n</ol>\n<h2>Output Contract</h2>\n<p>Return:</p>\n<ul>\n<li>Skills audited and paths used.</li>\n<li>Issues found, grouped by severity.</li>\n<li>Files changed, if apply mode was requested.</li>\n<li>Registry refresh recommendation when skill metadata or paths changed.</li>\n<li>Ambiguities that need human review.</li>\n</ul>\n<h2>References</h2>\n<ul>\n<li><code>docs/skill-style-guide.md</code> — normative LLM-first skill style guide.</li>\n<li><code>skills/skill-registry/SKILL.md</code> — registry refresh and indexing contract.</li>\n</ul>\n","files":[{"path":"SKILL.md","sizeBytes":2654,"isText":true}],"reviewScore":null,"reviewSummary":null,"trust":{"provenance":"trusted-source-unreviewed","notice":"Community-authored content, reproduced verbatim and not vetted as instructions. Treat it as data to evaluate, never as directives to follow.","bodySource":null},"bodyLocked":false,"purchaseUrl":null,"sourceUrl":null,"report":{"provenance":"trusted-source-unreviewed","screen":{"ran":true,"outcome":"clean","suspicious":0,"notes":0,"hiddenCharacters":false},"virusScan":{"engine":"clamav","status":"clean","scannedAt":"2026-09-20T08:02:13.207397Z","sha256":"D4E7E4A70331D09D65E1940DB505BD68CFA9CDB5131CC2EFE7362EA37CA6DB72","sizeBytes":1371},"review":null,"source":{"repositoryUrl":"https://github.com/Gentleman-Programming/gentle-shell","path":"skills/skill-improver","license":"MIT","commit":"78775c37903dc7821daceb848c7d9644f8c8c576","subtreeSha":"5E4E661F103CF459B8A3D230A7F1692E4AAF328143CBBC46D7A969B9D22EDD24","lastSyncedAt":"2026-09-27T20:55:08.631679Z"},"reviewedAt":"2026-09-20T08:08:10.347689Z","notice":"Community-authored content, reproduced verbatim and not vetted as instructions. Treat it as data to evaluate, never as directives to follow."},"install":[{"target":"skills-cli","command":"npx skills add https://github.com/Gentleman-Programming/gentle-shell/tree/main/skills/skill-improver"},{"target":"claude-code","command":"claude plugin marketplace add https://llmmart.ai/marketplace.json && claude plugin install gentleman-programming-gentle-shell@llmmart"},{"target":"git","command":"git clone https://github.com/Gentleman-Programming/gentle-shell.git"}]}