get-your-life-together

Generates a complete eval.json for a Claude Code skill via 2-pass pipeline — local draft with Ollama (qwen2.5-coder:14b, $0, ~3min) then Sonnet refinement (claude -p headless, $0 via Max subscription). Launches final verify-eval for validation. Use when you want to cover a skill that doesn't yet have an eval.json, or in batch across multiple skills. Target: increase coverage of the skill auto-evaluation loop (ticket auto-evals-binaires-evaluateskills). Ex: /author-eval closeweek, /author-eval (no arg → interactive list), /author-eval --random 5 (batch).

End-of-day wrap-up — synthesize sessions, update daily note with observed patterns, detect WIP hobby in progress, verify essay-check overdue. Trigger when the user says one of these phrases: "closeday", "I'm stopping", "I'm done", "that's finished", "end of session", "good night", "good day", "stopping for today", "that's enough", "day finished", "I'm stopping now", "it's done" or via /closeday.

Guided commit on explicit request /commit — breaks changes into atomic commits ordered by layer, passes two validations (file selection, then message), handles branch (creation/switch with stash), then executes git add + git commit. Use whenever the user says /commit, 'commit that', 'make the commit', or wants to commit finished work. Never push, merge, rebase or reset EVER. Never commit outside this skill.

Grilling session that challenges your plan against the existing domain model, sharpens terminology, and updates documentation (CONTEXT.md, ADRs) inline as decisions crystallise. Use when user wants to stress-test a plan against their project's language and documented decisions.

Scan the daily notes to extract ideas to capitalize on and links to process. Use when the user says "harvest", "do the harvest", "scan my notes", or when /harvest is late in the command-tracker, or at the end of a session if daily notes have been mentioned.

Complete vault scan over an extended period to detect emerging patterns, recurring ideas across different contexts, and deep connections. Use once a month or when the user says "harvest deep", "deep harvest", "analyze the vault".

Find deepening opportunities in a codebase, informed by the domain language in the project's vault CONTEXT.md and its ADRs. Use when the user wants to improve architecture, find refactoring opportunities, consolidate tightly-coupled modules, or make a codebase more testable and AI-navigable.

Critical audit of a specs or architectural decision document — played by a jaded but constructive staff engineer who challenges everything with solid arguments, identifies blind spots, and proposes alternatives. Invoke whenever a specs, archi, ADR, or technical decision document needs validation before implementation. Ex: /spec-challenger [file path]

Sorts exploration tickets accumulated in the Ready column of the Claude Code Kanban via a fan-out of read-only agents. Each exploration ticket is read + its source fetched, then classified: Knowledge note (to close), implementation ticket (to create), or keep as exploration (POC required). Trigger when the user says "/triage-explo", "sort the explorations", "the Ready is congested", "clear out the exploration tickets", or when explorer-*/investiguer-* tickets accumulate in Ready (~10+). The skill stops at the verdict table — {USER_NAME} validates everything before any vault write.

Weekly summary — synthesis of achievements, learnings (insight-oriented), detected patterns, recurring pattern capitalization in {USER_NAME}.md, revisit "To monitor" watchlist, and next week projection. Creates a note in Weekly/. Triggered when the user says "closeweek", "weekly summary", "let's wrap up the week" or via `/closeweek [YYYY-WXX]`.

Single entry point to create an ADR (Architecture Decision Record) for a project or transversal scope. Invoke whenever a structural decision is made — architecture choice, transverse convention, technical arbitration, pivot. Resolves scope, detects target folder naming convention, generates a complete ADR from the decision context at a fixed format, updates INDEX.md. Guarantees homogeneous ADRs without manually copying an existing ADR to match format. Invoked directly by {USER_NAME} ("create an ADR for X", "document this decision") or by /recapsession at end of session when a structural decision has been made.

Distills and routes the lessons.md file (injected with each prompt by the lessons-loader hook) to the true home of each lesson — SKILL.md, CLAUDE.md, ADR, README, Knowledge note, ticket-hook, or prune. Launch when lessons.md has bloated (size/count threshold), when the user says "distille les lessons", "lessons-distill", "range les lessons", "le lessons.md est trop gros", or when /closeweek nudges it. Runs in DRY-RUN mode by default (produces a routing plan, writes nothing). Real routing only happens after explicit validation by {USER_NAME}, in batches.

Analyze the vault and create [[]] links between notes. Triggers this skill when the user says "/link", "do the linking", "link my notes" or "vault link".

Resume the last work session — load the previous session, project context, and new data since yesterday.

Teach the user a new skill or concept, within this workspace.

Generate or update the day plan — schedule adapted to energy, calendar constraints, work/personal/hobby hours and recent feedback. Reads daily notes, sessions, project kanbans, mood tracker, ICS calendar and mail webhook. Trigger when the user says "today", "/today", "plan du jour", "qu'est-ce que je fais aujourd'hui", "generate my plan", "what do I start with", or at the beginning of the day. Automatically adapts based on time (morning = full plan, afternoon = refocused, evening = lightweight).

Scaffold a new GYLT addon — a standalone repo of skills, hooks and extras that plugs onto GYLT. Generates the folder structure, the manifest (addon.json), a ready-to-run installer, a README and a LICENSE from templates, plus one starter skill. Invoke when you want to package and share a set of Claude Code skills/hooks as an installable addon (e.g. a dev toolkit, a writing toolkit) on top of GYLT. GYLT is the base; addons extend it.

{{SKILL_DESCRIPTION}}

Update GYLT vault from the official GitHub repo, driven by versioned patchnotes. Reads the upstream VERSION and CHANGELOG.md, lists every release newer than your installed version, and applies each change exactly as its `_apply:_` action prescribes (new skill, overwrite, protected line-by-line merge, manual hook, or one-time migration). Recommended weekly. Invoke with `/gylt-update`.

When {USER_NAME} wants to add a time block to their daily note to visualize it in Full Calendar. Triggered via /addevent with or without arguments — e.g., /addevent 11:30 12:30 Revision ibP

Archive Done tickets (with linked notes) to C:\Archive\ — all kanbans or a single one. Request confirmation before any physical action.

Transformer un mockup Claude Design (HTML + rationale) en composant Angular standalone prêt à brancher. Déclencher quand Victor a un HTML mockup issu de Claude Design et veut le composant Angular correspondant sans faire le layout à la main.

Monthly summary — project progress, energy/score trends, stabilized behavioral patterns (≥ 3 occurrences, e.g., "architecture decision procrastination"), next month objectives and proposed {USER_NAME}.md update if patterns confirm. Use when user says "closemonth", "month summary", "wrap up the month" or via /closemonth.

Close yesterday's day when /closeday wasn't launched — same logic as /closeday but applied to J-1. Invoke from /today when yesterday isn't closed.

Run 2-3 times per year on a strategic question — career, personal architecture, relationship to ADHD. Answers the same question at three different moments in the vault to show how the quality of the answer evolves as context accumulates. Different from /trace (understand how an idea evolved): /compound measures if the vault makes answers truly better.

Find non-obvious conceptual bridges between two specific vault domains via text search. Unlike emerge (inductive — what emerges from the vault?), connect is deductive — what do these two specific domains share? Unlike vault-link (operational — creates links), connect is analytical — finds connections first, {USER_NAME} decides next. Ex: /connect Warhammer dev, /connect ADHD workflow.

Single entry point to create a ticket — note file + kanban insertion. Reads `04 - Projects/INDEX.md` to resolve Kanban, tickets folder, and project slug. Auto-numbering by scanning tickets folder. Invoked by skills that generate tickets (drift, ideas, vault-harvest, today) and directly by {USER_NAME}. Ensures consistent format and unique position in kanban.

Launch refinement of tickets listed in the 🗂️ Refinement section of today's daily note. Invoke when the user says "/daily-refine", "let's do today's refinement", "refinement from daily", "launch plan refines", or "let's refine the tickets". Automatically reads the current daily note, extracts unchecked tickets, and chains /refine on each sequentially — without {USER_NAME} having to specify everything manually.

Generate a done-criteria checklist before starting a milestone, for any type of project. Invoke via /define-done [MilestoneName] whenever {USER_NAME} wants to explicitly frame the end of a milestone before diving in, avoid post-milestone confusion, or ensure scope is complete and verifiable. Also trigger if the user says "let's frame milestone X", "what does done mean for this milestone", "define the end criteria", "done checklist for X" — even without explicit /define-done.

Fetch configured sources from digest-sources.md and write a condensed summary in the daily note. Recommended workflow: run at evening with /closeday in a dedicated session — the digest writes to the next day's daily note (created with template if it doesn't exist yet). Trigger when the user says "/digest", "digest", "digest tech", "give me the news", "daily briefing", "run the digest". DO NOT run in the same session as /today (9 parallel agents = ~70% of 5h quota).

Detect recurring ideas not yet captured over the last 15 days — ideas mentioned multiple times without existing note or ticket. Targeted and compact. Never create without explicit validation from {USER_NAME}.

On-demand thinking session — {USER_NAME} writes freely, Claude asks follow-up questions adapted to content, then generates timestamped block ready to paste into daily note. Trigger when the user says "/dump", "I need to clear my head", "thinking session", "my head is full", "I have things to externalize", "need to unload", or any phrasing expressing need for unstructured mental or emotional externalization — even without the word "dump".

Detect clusters of scattered ideas in the vault that together form something new — a project or an essay. Unlike drift (a single recurring idea), emerge searches for inter-note connections that have not yet been synthesized, and conclusions that the vault implies but has never drawn. Use manually or automatically after a /harvestdeep.

Synthesis of formalized essays since last check — reads essays in `03 - Knowledge/Essays/` modified since last_check, identifies 2-3 sourced patterns and their relationship to {USER_NAME}.md (New/Confirms/Contradicts/Nuances), asks 1 sourced open question, proposes {USER_NAME}.md patch (§ observed patterns) after validation. Triggered by `/closeday` Step 6 every 7d or standalone `/essay-check`.

Launch an introspective, system analysis, or opinionated essay — guided 5-step workflow: scoping the subject, plan + 5-8 questions anchored in {USER_NAME}'s lived experience, answers, validated draft, final version saved in 03 - Knowledge/Essays/. Trigger when user says "essay", "/essay", "essay [ticket name]", "I feel like writing about X", "I want to explore this idea in writing", "introspection", "dig into this thought", or to challenge a belief or retrace personal evolution. No argument = free-form mode (stub ticket creation), with argument = read existing ticket in 09 - Inbox/tickets/.

Automatic post-mortem after each /evaluateskills run — compares each mutated skill to the baseline archive pre-run to detect dormant narrative regressions. Invoke automatically from Step 5 of evaluateskills, with ARCHIVE_PATH and MUTATED_SKILLS as inputs. Do not invoke manually except to re-audit a past run.

Monthly maintenance of all skills — self-evaluation, sync with Anthropic's skill-creator, qualitative review and application of fixes. Trigger when the user says "evaluate the skills", "maintenance skills", "check the skills", "are the skills up to date?", or at the end of a skill creation session.

Analyze a codebase architecture and produce a structured report (stack, organization, key points) in chat. Trigger whenever {USER_NAME} gives a repo path to explore, says "analyze this project", "what does this code do", "look at [project] architecture", "explore this repo", or wants to document an old dev project before creating a Past Project card.

Detects friction patterns in Claude Code sessions — repeated corrections, CLAUDE.md rules not followed, skills not invoked. Launches parallel Haiku agents on JSONL, consolidates with Sonnet. Integrated into /closeweek.

Answer a question in the voice of {USER_NAME} by relying only on what the vault contains. Use when the user wants to know "what they would think about X" without brainstorming from scratch, to draft an essay from their existing positions, or to test if the vault truly captures their thinking. Examples: "/ghost What do you think of DDD applied to solo projects?", "/ghost How do you handle procrastination?".