Publish an SEO post to the WebJs Instagram account (@webjs_dev). Use this skill whenever the user asks to post to Instagram, publish an Instagram post, share something on Instagram, or promote a page/feature/blog on Instagram for SEO. Every post is strictly for SEO reach, so every post MUST ship a freshly created branded image plus a keyword-rich caption. The skill holds the account id, the credential location, the image recipe, the public-hosting requirement, and the two-step publish flow. It never prints or commits the access token.
Build and review WebJs applications. Use when working on WebJs app structure, pages, layouts, routes, server actions, components, signals, data and validation, auth, sessions, styling, the client router, streaming, or tests. WebJs is AI-first, web-components-first, and has no build step.
Use this skill whenever a change affects what `webjs create` GENERATES (a new or changed gallery/showcase demo, a new template, a changed generated file like the layout/home/theme/schema, a new convention that belongs in generated apps, a new scaffold-shipped config/hook) OR when the user asks to sync the scaffold, "update all three templates", check the scaffold is consistent, or teach agents how to use webjs through the scaffold. The scaffold is webjs's PRIMARY teaching surface for AI agents, so a change to it must propagate in lockstep across the generators, the per-agent rule files, the scaffold tests, the framework docs that describe the scaffold, and the preview/example apps. This skill is the authoritative map of every scaffold surface plus the change-type to surface mapping, and the mandatory "generate + boot + check" verification, so no surface is silently skipped. Complements webjs-doc-sync (which owns framework API/behaviour docs); this skill owns what the scaffold emits.
Use this skill when the user asks to start work on a tracked GitHub issue in the webjsdev/webjs project. Trigger phrases include "work on
Use this skill whenever the task is to write, draft, or substantially edit a webjs blog post (a marketing/engineering post under the repo-root `blog/` directory that the website publishes). It carries the author's voice, the hard prose rules (no em-dashes, no internal PR/issue numbers, no process tells), the SEO-first topic and front-matter conventions, the de-duplication check against every existing post, and the mandatory dogfood verification of every factual claim before a feature post ships. Invoke it BEFORE writing any blog prose so the post lands in the house voice and does not need a later cleanup pass.
Use this skill whenever a change ships a user-facing or agent-facing surface (a new export, CLI flag, package.json webjs.* config key, html hole prefix, lifecycle hook, convention, or a behaviour change to an already-documented feature) and the docs must be brought in sync, OR when the user asks to find documentation drift / doc gaps / "did we update the docs", audit shipped work for missing docs, or sync the docs surfaces. The skill carries the authoritative map of EVERY doc surface webjs ships and the change-type to surface mapping, so no surface (the docs site, the marketing website, the skill at `.agents/skills/webjs/`, README, AGENTS.md, the scaffold templates' per-agent rule files) is silently skipped.
Use this skill to record a research, design, or decision investigation on the webjsdev/webjs project. A research/design record is a CLOSED GitHub issue labeled `research` (the durable writeup lives in the issue body plus comments), never a file in the framework's reference docs (the skill at `.agents/skills/webjs/`) and never a comment buried on an unrelated PR. Trigger phrases include "research whether X", "investigate X", "evaluate X vs Y", "compare A and B and decide", "design record for Z", "decision record", "write up the design", "spike X", or any natural-language ask whose deliverable is a writeup of a comparison, options, or a decision rather than shipped code.
Operate Railway infrastructure: create projects, provision services and databases, manage object storage buckets, deploy code, configure environments and variables, manage domains, troubleshoot failures, check status and metrics, and query Railway docs. Use this skill whenever the user mentions Railway, deployments, services, environments, buckets, object storage, build failures, or infrastructure operations, even if they don't say "Railway" explicitly.