l1veIn
UserYour repo, but make it a girl! 🎨 Mascots, icons, stickers, posters & landing pages for your repo — driven by your own coding agent. No embedded runtime.
Categories
Indexed Skills (9)
repochan-analysis
Analyst role — deep analysis with LLM enrichment. Runs a deterministic scan (Steps 1–5, 7), then performs LLM pre-analysis (Step 6) and abstract dimension analysis (Step 8), finally writing the analysis report (repochan analysis). Use when analyzing a repo, running repochan analysis, or when the user asks to analyze a repo / scan a project / produce an analysis report.
repochan-art-director
Art Director & Product Manager role. Creates all Asset Orders at once (foundation + downstream), ensuring character consistency. Downstream orders reference the foundation, and the Painter executes in dependency order. Use when creating asset orders, foundation sheets, template curation, or when the user asks about Art Director / Asset Orders / creation tasks / foundation sheet tasks.
repochan-interviewer
Interviewer role. In standard mode: reads the analysis report, asks the user 7-14 structured questions across 8 dimensions (tone, audience/usage, weight, world, style, reference, naming, constraints) via ask_user_question, then distills answers into an interview report for the Creative Team. In greenfield mode (no existing repo): extracts project intent from the user's vision using greenfield-specific dimensions (project essence, target audience, tone, naming, visual style, scope, tech, constraints). Use when running interviews, repochan interview create/append, or when the user asks about interview/questionnaire/preference collection.
repochan-page-designer
RepoChan Starter Localizer. Select and pull a pre-designed complete Starter, apply project configuration, rewrite full locale, create and apply slot assets, run localization validation. Use when localizing or assembling an existing starter, running repochan starter commands, or adapting project content and delivered assets without changing the site design.
repochan-painter
Painter role. Execute Asset Orders assigned to the Painter (auto-approve draft when appropriate): resolve visual references, prepare briefs, select the best image generation path, save order result versions. Execute the foundation-sheet-first consistency model. Use when painting/generating images for asset orders, running repochan image gen, handling order review/revision/candidate, or when the user asks painter/generate image/regenerate/create image.
repochan-persona
Creative Team role. Uses three-agent collaboration (World Architect + Character Designer + Consistency Guardian) to generate a living mascot persona from repo analysis + optional interview report. Supports grounded/high-concept characters, strictly following project-weight matching, anti-overfit, anti-language-leak, and world-character synergy principles. Use when generating/revising mascot personas, running persona create/update/review/candidate, or when the user asks for persona/character/mascot/persona design.
repochan-starter-designer
RepoChan Starter Productization Engineer. Take a complete project website that has passed Gate 2 and faithfully organize it into a creator-owned, localizable, verifiable, PR-contributable source starter. Use when productizing an approved implemented page into a creator-owned starter, defining localization slots and migration references, or preparing an official starter contribution.
repochan-web-designer
RepoChan original web designer. For a specific git project, complete from scratch: website art direction, full-page and section visual master design, bake-mask, production asset strategy, responsive implementation, and Gate 1/2 acceptance. Use when a project needs an original website, a new section or information architecture, a new art direction, image-driven page construction, or when no existing starter fits.
repochan
RepoChan Wizard — turn a git repo into a complete brand asset suite (persona, illustrations, stickers, landing page) and prepare for deployment. Default is guided mode: run the full pipeline stage by stage, stopping at 3 checkpoints for user confirmation — do not run to completion unprompted. Only when the user explicitly says "yolo" do you adopt default creative decisions within the authorized scope without stopping; external writes like deployment still require explicit authorization, and non-interactive environments do not expand authorization. Per-team access is the advanced mode. Use when the user runs /repochan, wants the full pipeline, or says "one-shot" / "full pipeline" / "yolo".
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.