Development

Build or bulk-edit an Open edX course outline on Studio — a new course in one request from a JSON spec via the IBL course-creator API (POST /api/v1/ibl/course/create_full_course, which also publishes), or an existing course at scale with the bundled reconcile driver that diffs the same spec against the live outline and creates/updates/reorders/deletes idempotently; plus the v1 helpers to list every unit with its components and create/update/publish single blocks. Use when a course plan is kno...

Verify and operate the LMS side of an Open edX course built on Studio — the calls the lms.ibl.ai LMS app makes against the LMS API: enroll yourself or (as course staff) other users, check enrollment status, list a course's users, read the published outline and blocks as a user, read course details and the merged course metadata the catalog shows, and progress/grading as seen by users. Use after publishing to confirm what users see, to enroll test users, or when the user asks "is it in the cat...

Add and edit HTML (text) components in an Open edX unit on Studio — create the html block, write its HTML body and title in one update, reference uploaded images and files via /static/ paths, duplicate or delete it, and follow the authoring rules that make content look designed (cards, callouts, tables, tooltips via inline CSS) while leaving typography and page width to the LMS. Use when the user wants text, a reading, an explanation, a callout, a comparison table, an embedded video/iframe, o...

Create a course on Open edX Studio (studio.learn.iblai.app) from the terminal — POST the org key and display name (optionally number and run) to Studio's course-management endpoint and get the course key back; list the user's courses, read a course summary, derive the course root block for outline work, and delete a course. Use when the user wants a new Open edX course, asks for their course key, needs to know which courses they can edit, or wants a course removed. Session auth via studio.env...

Sign in to Open edX Studio (studio.learn.iblai.app) and the LMS from a browser window the user controls, capture the session cookies (studio_session_id, sessionid, csrftoken) with the bundled Playwright script, verify them, and write the gitignored studio.env that every /iblai-api-studio-* skill reads; --check validates an existing session in one command. Use first whenever the user wants to author or configure an Open edX course from the terminal, and again when a Studio call answers 302 to ...

Manage an ibl.ai organization's learning catalog via the platform API — courses, programs, pathways, resources, skills, roles, course/program metadata, plus enrollment, eligibility checks, catalog search, and course reviews. Org-wide content and enrollment operations. Use when wiring up the catalog, enrolling users, checking eligibility, or curating skills/roles/pathways.

Build and populate Canvas LMS courses through the Canvas REST API — creating course shells, modules and module items, pages, assignments, quizzes (Classic and New Quizzes), discussions, announcements, files, assignment groups, sections and enrollments, plus bulk seeding from a spec file, course copies and Common Cartridge imports. Use this skill whenever the user mentions Canvas, Instructure, canvas.instructure.com, a Canvas API token, `/api/v1/courses`, or asks to script, seed, migrate, bulk...

Decide how money works for an ibl.ai app — how your organization is charged (prepaid credits, plans, auto-recharge, your own LLM keys, spend caps) and which of the three rails to use to charge your own users (platform credits, an app paywall on your own Stripe key, or item-level monetization via Stripe Connect). Use when the user mentions pricing, charging users, paywall, subscription, credits, "how do I get paid", Stripe, spend limits, budgets, or LLM cost. For the credit widget see /iblai-v...

Add the OS app's Code mode to a vibe Next.js + Tauri app — a "Code" pill in the chat composer that turns the agent into an agentic coding tool (opencode over ACP) editing files, running commands, and committing in a workspace on the user's machine, with per-operation Allow/Deny permission cards, a loopback proxy that never hands the agent the real token, per-agent skills synced from the platform, and phone access by pairing a mobile build to a desktop. Ships the Tauri Rust modules from src-ta...

MCP connectors for ibl.ai agents, end to end — register MCP servers (featured cross-organization sharing, enable/disable, OAuth service linkage), manage credential connections with the auth_type × scope matrix (org / agent / per-user), wire servers onto an agent, run the OAuth connected-service lifecycle, handle per-user in-chat OAuth (the oauth_required / oauth_connection_resolved chat events), and debug MCP auth failures (401s, missing OAuth prompts, agents ignoring servers). Use when wirin...

Safely manage registered project environment variables through Kavranta's redacted local broker. Use for `.env` or Wrangler `.dev.vars` inspection and edits, groups, access policies, links, value-hidden generation or reuse, deployment and comparison, encrypted team sharing, and Provider or Action Packs. 한국어로 “환경변수 관리·정리”, “.env를 값 없이 확인”, “키 연결·재사용”, “환경변수 올리기”, “카브란타로 처리”를 요청할 때도 사용한다.

Establish Specs from an explicitly selected existing implementation, then retire this temporary adoption Skill after successful finalization.

Status: Accepted Supersedes the `sb-discovery` ownership and routing portions of Decisions [0181](./0181-reverse-spec-establishment.md), [0188](./0188-retire-legacy-staged-adoption.md), [0191](./0191-resumable-reverse-establishment.md), and [0192](./0192-typed-milestone-action-handlers.md), and [0193](./0193-progressive-discovery-procedures.md). Their fixed-evidence, confirmation, lifecycle, resume, and non-release finalization contracts remain authoritative.

Review and integrate Dependabot pull requests in Huruikagi/specbind one at a time, with current-branch checks, main CI verification, and affected dependency or toolchain validation. Use when asked to handle or merge SpecBind Dependabot PRs; do not use for dependency upgrades that have no Dependabot PR.

MANDATORY for every coding agent and contributor touching localized content — keep all five localization surfaces (dashboard UI keys, wiki page, mirrored READMEs, locale-aware formatting, language switchers) in parity across every supported language. Use automatically (without being asked) whenever you add or change user-visible UI copy, add an i18n key, edit README.md, edit wiki/index.html, or change docs that the READMEs and wiki mirror — and use the full new-language checklist whenever add...

MANDATORY for every coding agent and contributor touching localized content — keep all five localization surfaces (dashboard UI keys, wiki page, mirrored READMEs, locale-aware formatting, language switchers) in parity across every supported language. Use automatically (without being asked) whenever you add or change user-visible UI copy, add an i18n key, edit README.md, edit wiki/index.html, or change docs that the READMEs and wiki mirror — and use the full new-language checklist whenever add...

MANDATORY for every coding agent and contributor touching localized content — keep all five localization surfaces (dashboard UI keys, wiki page, mirrored READMEs, locale-aware formatting, language switchers) in parity across every supported language. Use automatically (without being asked) whenever you add or change user-visible UI copy, add an i18n key, edit README.md, edit wiki/index.html, or change docs that the READMEs and wiki mirror — and use the full new-language checklist whenever add...

What the gates cannot tell you about adding or changing a zh-TW rule - why assets/ruleset.json is the only place a vocabulary rule lives, the false-friend problem and the four gates that answer it, the corpus thresholds a rule has to clear, why positions are byte offsets through NFC, and how a rule reaches the scanner, the fixer and the browser build. Use when adding or disabling a rule, when a rule fires on native zh-TW prose, when a fix lands at the wrong offset, or when touching src/engine...

Use when implementing any feature or bugfix, before writing implementation code

Behavioral guidelines to reduce common LLM coding mistakes. Use when writing, reviewing, or refactoring code to avoid overcomplication, make surgical changes, surface assumptions, and define verifiable success criteria.

Anti-slop frontend skill for landing pages, portfolios, and redesigns. The agent reads the brief, infers the right design direction, and ships interfaces that do not look templated. Real design systems when applicable, audit-first on redesigns, strict pre-flight check.

Write what GitHub itself reads: README, CONTRIBUTING, the community-health files (SECURITY, CODE_OF_CONDUCT, LICENSE, issue and PR templates) and the workflows under .github/workflows/. Use when creating or rewriting any of these, when Community Standards reports a gap, when adding or reviewing CI, or when a repository goes public. Not for docs/, which is writing-docs.

Find a Claude Code skill or plugin somebody else has already written, and judge whether to install it. Use before writing a skill for something that sounds common, when asked whether an integration exists, when looking for a plugin by capability rather than name, or before adding a skill that would duplicate one installed.

Firmware development for the Guition JC-ESP32P4-M3-DEV board (JC-ESP32P4-M3 module — ESP32-P4NRW32 with 32 MB in-package PSRAM, 16 MB flash, and an ESP32-C6 radio co-processor on SDIO) — its MIPI-DSI panel and MIPI-CSI camera connectors, IP101 Ethernet, ES8311 codec with speaker amplifier and microphone, microSD on SDMMC, RS485, three USB-C sockets, the 26-pin JP1 header, and the PlatformIO + ESP-IDF (pioarduino) setup around them. Use when working on this board or any ESP32-P4 devkit: projec...