Catalogue des skills
Les skills ShipGlows découpent le modèle opérationnel en actions réutilisables : trier un bug, cadrer une spec, vérifier la préparation, auditer un système ou rechercher un marché avant de perdre du temps de build. Le but n’est pas plus d’astuce de prompt. Le but est moins de passages de relais faibles.
Langue des skills
C’est normal : ces fiches décrivent des workflows destinés aux agents, et les agents les exploitent plus sûrement en anglais. ShipGlows garde donc l’interface de navigation en français, tout en conservant les contrats de skills dans leur langue de travail.
Skills mis en avant
Start with one plain instruction and let ShipGlows choose the right workflow.
The primary non-technical router for answering questions, executing deterministic micro-edits directly, or handing substantive work to the right ShipGlows skill.
Open skill pageRun non-trivial work from story to spec, build, verification, closeout, and ship without making the user drive every gate.
The master user-facing lifecycle orchestrator for carrying a story, bug, or goal through ShipGlows's spec, readiness, implementation, verification, onboarding consideration, documentation, closure, and shipping gates.
Open skill pageTriage a bug fast, decide whether it is safe to patch now, and route it correctly.
The bug-first entrypoint for ShipGlows when the work starts from broken behavior instead of a new feature request.
Open skill pageTurn a fuzzy request into an implementation-ready contract before code starts mutating the repo.
The spec-writing skill for shaping non-trivial work into a clear user story, behavior contract, executable task list, and chantier history.
Open skill pageStart implementation when the task is ready, and force a stronger path when it is not.
The main execution entrypoint for beginning real work after the problem and contract are clear enough, while keeping the chantier flow visible.
Open skill pageCheck whether the work really satisfies the user story, not just whether the code compiles.
A verification skill for standard readiness or an explicit excellence-focused second pass beyond the task's acceptance criteria.
Open skill pageGuide the human through the real user flow, then turn the result into test evidence and actionable bugs.
A manual QA skill for creating guided test campaigns, capturing structured outcomes, and preserving bug history after implementation.
Open skill pageRun a broad project audit when you need a high-level map of what is weak before you decide what to fix.
The top-level audit skill scans across multiple quality domains so you can see where the real risk or drift sits.
Open skill pageManage content work through one lifecycle instead of guessing which writing, audit, docs, or SEO skill to launch.
A master content lifecycle skill for routing strategy, repurposing, drafting, enrichment, audits, public-content governance, validation, and ship readiness through the right owner skills.
Open skill pageUpgrade content with fresher research, stronger structure, and more useful answers.
A content-improvement skill for taking an existing page or article and making it more current, more actionable, and more discoverable.
Open skill pageChoose the right marketing review without choosing among four overlapping skills.
One marketing skill with explicit modes for market evidence, GTM readiness, copy clarity, and persuasion.
Open skill pageOne design skill, with explicit modes and rigorous playbooks.
The single public design entrypoint for system creation, playgrounds, UI/token/component/accessibility audits, implementation, proof, verification, and shipping.
Open skill pageTurn shipped features into clear, trusted customer journeys.
A customer-experience skill for UX/UI clarity, friction, trust, first-success paths, onboarding, setup sequencing, recoverable states, docs coherence, and proof routing.
Open skill pageDesign local-to-cloud sync without losing user data.
A data-sync skill for local data promotion, cloud hydration, merge/conflict policy, sync UX states, sensitive-data exclusions, and proof routing.
Open skill pageCatalogue complet
Start with one plain instruction and let ShipGlows choose the right workflow.
The primary non-technical router for answering questions, executing deterministic micro-edits directly, or handing substantive work to the right ShipGlows skill.
Open skill pageExplore a problem space before coding so you do not harden the wrong idea into implementation.
A non-coding exploration skill for clarifying ambiguity, options, and risks before the build phase starts.
Open skill pageRun a bug from report to dossier, fix, retest, verification, and ship risk without skipping gates.
A professional bug loop executor that reads bug state and continues through the right ShipGlows owner skills when safe.
Open skill pageRun non-trivial work from story to spec, build, verification, closeout, and ship without making the user drive every gate.
The master user-facing lifecycle orchestrator for carrying a story, bug, or goal through ShipGlows's spec, readiness, implementation, verification, onboarding consideration, documentation, closure, and shipping gates.
Open skill pageTriage a bug fast, decide whether it is safe to patch now, and route it correctly.
The bug-first entrypoint for ShipGlows when the work starts from broken behavior instead of a new feature request.
Open skill pageTurn a fuzzy request into an implementation-ready contract before code starts mutating the repo.
The spec-writing skill for shaping non-trivial work into a clear user story, behavior contract, executable task list, and chantier history.
Open skill pagePressure-test a spec before implementation so the thread is not coding against loose intent.
A readiness gate for checking whether a spec is complete, safe, and executable by a fresh agent.
Open skill pageStart implementation when the task is ready, and force a stronger path when it is not.
The main execution entrypoint for beginning real work after the problem and contract are clear enough, while keeping the chantier flow visible.
Open skill pageCheck whether the work really satisfies the user story, not just whether the code compiles.
A verification skill for standard readiness or an explicit excellence-focused second pass beyond the task's acceptance criteria.
Open skill pageGuide the human through the real user flow, then turn the result into test evidence and actionable bugs.
A manual QA skill for creating guided test campaigns, capturing structured outcomes, and preserving bug history after implementation.
Open skill pageKeep product expectations and technical implementation aligned across every supported platform.
Audits platform parity, separates legitimate platform adaptations from implementation debt, and routes missing work to the right ShipGlows skill.
Open skill pageClose a task cleanly so the work is summarized, tracked, and ready for the next thread.
A task-closing skill for wrapping up implementation, documenting completion, and recording closure in the chantier flow when applicable.
Open skill pageReview recent work and the session state before you lose the thread of what changed.
A review skill for summarizing recent changes, identifying remaining gaps, and updating local workflow state cleanly.
Open skill pageCommit and push quickly when the work is actually ready, instead of stretching closure into another vague step.
A shipping skill for moving finished work through the final commit-and-push path and recording the shipping result when a chantier spec is in scope.
Open skill pageRun a broad project audit when you need a high-level map of what is weak before you decide what to fix.
The top-level audit skill scans across multiple quality domains so you can see where the real risk or drift sits.
Open skill pageUse one explicit technical mode for code risk, dependencies, performance, or a breaking migration.
A unified technical entrypoint with four bounded modes and evidence limits that match the work.
Open skill pageCheck whether the site is set up to earn search visibility in the current AI and SEO environment.
A technical and on-page SEO audit for discoverability, crawlability, content structure, and AI-era visibility signals.
Open skill pageReview whether translations are complete, consistent, and trustworthy, then sync missing localized content when the mapping is clear.
A translation audit and controlled i18n sync skill for completeness, tone consistency, cross-language coherence, and missing localized entries.
Open skill pageCapture ideas without letting them hijack the current execution thread.
A local-first backlog management skill for parking non-urgent work while keeping active delivery focused.
Open skill pageTurn recent git history into a human-readable changelog instead of leaving progress trapped in commits.
A changelog generation skill for summarizing meaningful project changes in a cleaner public format.
Open skill pagePrime the relevant files and decisions before the thread starts guessing about the codebase.
A context-loading skill that retrieves the most relevant repo surface before implementation begins.
Open skill pageCreate or repair documentation so the repo stays navigable for both humans and agents.
A documentation skill for maintaining docs, metadata, and governance-layout compliance across project surfaces.
Open skill pageManage content work through one lifecycle instead of guessing which writing, audit, docs, or SEO skill to launch.
A master content lifecycle skill for routing strategy, repurposing, drafting, enrichment, audits, public-content governance, validation, and ship readiness through the right owner skills.
Open skill pageClean up git and GitHub drift before it turns into release friction.
A git/GitHub hygiene skill for sync drift, stale branches, PR drift, and Dependabot backlog triage with bounded safe fixes.
Open skill pageBootstrap a project into the ShipGlows operating model instead of improvising the structure from scratch.
An initialization skill for detecting project shape, generating baseline governance docs, and registering the repo in the ShipGlows workflow.
Open skill pageKeep the task list coherent after work moves, instead of letting tracking fall behind reality.
A task-management skill for updating `TASKS.md`, reflecting progress, and capturing the next sensible steps.
Open skill pageUpgrade content with fresher research, stronger structure, and more useful answers.
A content-improvement skill for taking an existing page or article and making it more current, more actionable, and more discoverable.
Open skill pageChoose the right marketing review without choosing among four overlapping skills.
One marketing skill with explicit modes for market evidence, GTM readiness, copy clarity, and persuasion.
Open skill pageReorder the backlog by impact and effort before momentum turns into scattered activity.
A local-first prioritization skill for sorting active and pending work with a more explicit impact-versus-effort lens.
Open skill pageWrite original long-form content that sounds intentional, not generic.
A writing skill for producing long-form content aligned to the project voice, audience, and business goal.
Open skill pageResearch a topic deeply enough that you can make a better product or GTM decision afterward.
A structured research skill for gathering evidence, checking current-source freshness, and saving reusable workflow reports.
Open skill pageTriage external links and incoming information before they dissolve into random browser tabs.
A governance-aware link-analysis skill for reviewing URLs or pasted content, summarizing relevance, and routing follow-up actions.
Open skill pageRun the standard technical checks and fix the obvious issues before you pretend the work is done.
A practical quality gate for typecheck, lint, build, and related technical validation in the current project.
Open skill pageOne design skill, with explicit modes and rigorous playbooks.
The single public design entrypoint for system creation, playgrounds, UI/token/component/accessibility audits, implementation, proof, verification, and shipping.
Open skill pageRun the whole release confidence loop instead of stitching checks, ship, prod, browser proof, and verification by hand.
A release orchestration skill for moving work through technical checks, bounded shipping, deployment readiness, post-deploy evidence, final verification, and optional changelog routing.
Open skill pageTurn shipped features into clear, trusted customer journeys.
A customer-experience skill for UX/UI clarity, friction, trust, first-success paths, onboarding, setup sequencing, recoverable states, docs coherence, and proof routing.
Open skill pageRun project maintenance from triage through delegated fixes, verification, and ship routing.
A master maintenance lifecycle skill for reviewing project health, creating or continuing specs when needed, delegating remediation, verifying results, and routing ship or deploy.
Open skill pageDesign local-to-cloud sync without losing user data.
A data-sync skill for local data promotion, cloud hydration, merge/conflict policy, sync UX states, sensitive-data exclusions, and proof routing.
Open skill pageDesign product entitlement ownership and server-side access gates before premium behavior changes.
A product-entitlement skill for suite and standalone access models, provider events, redemption codes, support diagnostics, backend authorization, and sync handoffs.
Open skill pageDebug a broken auth flow in a real browser before you start guessing from static code alone.
A browser-level auth diagnosis skill for reproducing login failures, reading local Clerk/Google/Convex/Flutter references, and isolating the exact break point in the flow.
Open skill pageOpen a URL in a real browser and verify one observable non-auth objective.
A general browser verification skill for checking public UI, previews, production pages, visible assertions, screenshots, console messages, and network signals without turning every page check into auth debugging or manual QA.
Open skill pageCheck whether production is actually healthy after a push instead of assuming the deploy tells the truth.
A production verification skill for deployment status, logs, and live health checks after release.
Open skill pageAudit exported ShipGlows transcripts for systemic agent flaws and route durable fixes.
A focused audit lane that reads canonical conversation evidence, classifies repeatable failure patterns, and routes recommended owners.
Open skill pageGenerate new files that match the repo patterns instead of starting from a blank page every time.
A scaffolding skill for creating professional pages, components, layouts, routes, hooks, and utilities that fit the project shape.
Open skill pageGet a fast git-based project dashboard instead of manually checking every repo by hand.
A read-only status skill for seeing branches, changes, sync state, and recent activity across the current project or workspace.
Open skill pageName or rename the current session so the thread stays legible as work branches.
A tiny utility skill that keeps long-running agent work easier to scan, search, and resume.
Open skill pageSee the workflow map when you need to remember which skill to use next.
A reference skill that explains the ShipGlows system, available skills, and the intended execution paths between them.
Open skill pageChoose the right model for the task without letting speed or cost outrank quality.
A model-selection skill for matching task type, reliability, quality bar, latency, cost, and subagent routing needs before execution starts.
Open skill pageSummarize the current thread before context decay forces you to reread everything.
A fast recap skill for compressing the current conversation into the key points, task states, visible commits, and next-step reality.
Open skill page