{ "markdown": "# livlign skills\n\nA Claude Code plugin marketplace for open-source maintainers — README heroes, README audits, and trustworthy .NET test-coverage backfills.\n\n```\n/plugin marketplace add livlign/claude-skills\n```\n\n> Already added the marketplace? Run `/plugin marketplace update livlign` to pick up newly-published plugins (e.g. `dotnet-coverage-kit`) and version bumps.\n\n## Plugins\n\n| Plugin | Install | What it does |\n|---|---|---|\n| [repo-visuals](./plugins/repo-visuals) | `/plugin install repo-visuals@livlign` | Designs bespoke README hero visuals — animated GIF, static PNG, or animated SVG — through a structured discovery dialog. |\n| [repo-doctor](./plugins/repo-doctor) | `/plugin install repo-doctor@livlign` | Audits a repo's README against best-practice patterns and produces a prioritized P0/P1/P2 punch list of fixes. |\n| [dotnet-coverage-kit](./plugins/dotnet-coverage-kit) | `/plugin install dotnet-coverage-kit@livlign` | Backfills characterization + spec unit tests into .NET services and keeps C0/C1 coverage honest via a checked-in manifest and a CI gate. |\n\n---\n\n## `repo-visuals`\n\nScans the target repo, recommends a format that fits its identity, proposes scenarios, designs a bespoke HTML stage, previews it in your browser, and exports a retina-quality artifact — then optionally opens the upstream PR.\n\nFeatured in [**rohitg00/awesome-claude-code-toolkit**](https://github.com/rohitg00/awesome-claude-code-toolkit) — the curated list of Claude Code plugins, agents, and MCP servers.\n\n**What a run looks like:**\n\n```\n> /repo-visuals\n pick mode → semi-auto · scan repo → 30 tools detected\n recommend → animated GIF, 1200×675 · brief locked\n build HTML → preview in browser → ship\n export → hero.gif (2.4 MB) · open PR #42\n```\n\n**Why it's different**\n\n- **Discovery dialog, not templates.** Every hero is bespoke. A structured scan-then-design conversation runs before any HTML is written.\n- **Production-grade by default.** Retina capture, drift-proof copy (no stale star counts or version numbers), preview-vs-export parity, scope-match guardrails — encoded from real maintainer pushback.\n- **Three modes.** Manual, Semi-auto (recommended), Auto.\n- **Optional PR.** Local-only or full upstream PR with relative-path embed, alt-text generation, and fork-and-PR for repos you don't own.\n\n### Recent heroes shipped\n\nThumbnails — click through to see the live hero in each repo.\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
\n htmlhint/HTMLHint\n \"HTMLHint\n\n gui-cs/Terminal.Gui\n \"Terminal.Gui\n
\n kunwar-shah/claudex\n \"claudex\n\n rullerzhou-afk/clawd-on-desk\n \"clawd-on-desk\n
\n affaan-m/everything-claude-code\n \"everything-claude-code\n\n emtyty/ast-graph\n \"ast-graph\n
\n\n[Full showcase →](./plugins/repo-visuals/SHOWCASE.md)\n\n[Skill source →](./plugins/repo-visuals)\n\n---\n\n## `repo-doctor`\n\nAudits a README the way a senior maintainer would — does the \"what is this\" sentence land, is install-to-first-success short enough, does the audience-jargon match, are there drift signals (stale versions, badge sprawl, dead links). Read-only diagnostic by default; opens a PR only on explicit ask.\n\n**What a run looks like:**\n\n```\n> /readme-doctor\n scan repo → classify category (CLI tool · library · framework · …)\n audit → score 9 criteria with evidence\n output → P0/P1/P2 punch list, each with a one-line fix\n health score: 69/100\n```\n\n[Skill source →](./plugins/repo-doctor)\n\n---\n\n## `dotnet-coverage-kit`\n\nBackfills unit tests into legacy .NET services and keeps coverage honest. It classifies every file **per method** — so a controller or a DbContext-bound god-class still contributes its testable slice instead of being written off wholesale — writes characterization tests for old code and spec tests for new, and measures deterministic C0/C1 via dotnet-coverage. A checked-in manifest explains why every file is tested, excluded, or can't-be-tested (each with a mitigation), and a CI gate (diff coverage + ratchet + scope guard) stops it slipping.\n\n**What a run looks like:**\n\n```\n> /coverage-init\n detect projects → parallel sweep classifies every file (per-method carve-outs)\n draft manifest → self-critique loop → you confirm only the gray areas\n> /generate-tests\n characterize old code · spec new code → 100% of testable branches\n genuinely untestable → manifest entry with a reason + mitigation\n> /coverage-report\n deterministic C0/C1 → merge sets the baseline → CI gate on every PR\n```\n\n[Plugin source →](./plugins/dotnet-coverage-kit)\n\n---\n\n## Contributing\n\nEasiest entry points: report a real failure as an issue, pick something from the [wishlist](./CONTRIBUTING.md#3-pick-something-from-the-wishlist), or submit a hero to the [showcase](./plugins/repo-visuals/SHOWCASE.md).\n\nFull guide → [CONTRIBUTING.md](./CONTRIBUTING.md)\n\nPersonal project, evolving fast — expect breaking changes pre-1.0.\n", "bytes": 6175, "sha": "34807ca30762686fbb6b0699bbe0bd197f97859aa5da79fc9c1f98efd7fdb600", "repo_slug": "livlign/claude-skills", "fonte": "repo", "truncated": false, "api": "https://agentalog.com/api/listings/plg_livlign_claude_skills_repo_visuals_30f14762/readme" }