Skill Claude Code
Angular Core UI executor for approved implementation in new or existing @sdcorejs/angular and @sd-angular/core apps, covering portal init, CRUD/list/detail/form/action scope, explicit technical prototypes, or approved migration. Optional admin/auth packs require approved scope. Do not use for plain Angular apps or…
Skill Claude Code
Conditional architecture gate between approved spec and plan. Use for cross-module/repository boundaries, public/event/data contracts, state ownership, trust boundaries, major dependency or paradigm choices, or dependency direction before fan-out. Route mapping to explore, audits to review, open-ended design to…
Skill Claude Code
Discovery and requirement-confirmation gate before spec. Use for open-ended ideas, compare approaches, missing blockers, or CRUD/entity/screen/module/backend/site/AI-agent/test/product/design/doc/workflow requests across Angular, NestJS, Next.js, React, Node, fullstack, AI-agent, documentation, workflow, test…
Skill Claude Code
Systematic debugging skill for root-cause/fix of concrete bugs and failing tests. Use for debug, investigate, resolve, fix, root-cause, wrong behavior, flaky/CI/prod-only failure, runtime/build error, or performance anomaly. Use sdcorejs-test for write/run/plan tests without fix intent. Runtime-localized.
Skill Claude Code
Design-track executor for UI/UX design and improvements, mobile app/mobile web design, wireframes, mockups, screen flows, PNG previews, and frontend handoff from product stories. Use for new or existing UI design; independent audits/findings belong to sdcorejs-review and production code to the implementation executor.…
Skill Claude Code
Documentation executor/gate for documentation artifacts. Use for code comments/docstrings/JSDoc/TSDoc, user guides/manuals with verified UI evidence, TASKID or requirement records, technical/API docs, doc rewrite/improve/structure/summarize/convert. Do not use for project summary/code maps/env setup/recovery, product…
Skill Claude Code
Execute an approved plan snapshot. Use after sdcorejs-plan approval or when asked to execute, run, generate from, or continue an approved plan for AI-agent, Angular, NestJS, Next.js, React, Node, fullstack, product, design, documentation, workflow, test, or generic/general work. Detects track/profile, preserves…
Skill Claude Code needs its repo
Project exploration and durable context skill. Use to understand/map a codebase, trace feature flows, find reusable components/services, refresh .sdcorejs/summary.md, recover context, set/read persona, save/read memories, harvest documentation facts, or inspect local env shape. Modes: summary-read, summary-refresh…
Skill Claude Code
Git artifact workflow. Use for commit/save changes, create PR, push verified branch, changelog, release notes, diff since tag, or PR text. An explicit worktree request is a compatibility handoff to the orchestration workspace-isolation contract, not a Git artifact mode. Do not use for readiness/ship gates; use…
Skill Claude Code
NestJS code executor for approved/direct backend work with confirmed requirements. Use after execute-plan or for project/admin/module/entity CRUD, explicit endpoints, workflows, bulk operations, exports, or reuse of @sdcorejs/utils. Loads the canonical NestJS manifest and only the packs required for the resolved…
Skill Claude Code needs its repo
Next.js website code executor for approved site/module work with confirmed requirements and an explicit feature set. Use for approved bootstrap, theme, pages/sections, SEO/OG, i18n, caching/ISR, responsive, contact, analytics, CMS, or content-quality scope. Basic sites do not imply i18n or contact. Loads only approved…
Skill Claude Code
Low-level parallel safety scheduler. Use when sdcorejs-subagent-driven-development or execute-plan requests an attested wave verdict/fan-in, or when the user explicitly requests a read-only parallel review or audit. Validates protocol-v2 contracts, dependency waves, capabilities, path/resource ownership, task briefs…
Skill Claude Code
Plan authoring and approval gate. Use after approved spec, or when the user asks to write/review/approve/change an implementation plan. Produces numbered file/task plans, self-reviews, waits for explicit approval, then persists the approved snapshot in the matching track directory under .sdcorejs/plans/ before…
Skill Claude Code
Product-track executor for PO docs, traceability, and task breakdown. Use for PRDs, user stories, acceptance criteria, feature ledgers, requirement review, traceability matrix, UAT checklists, requirement/implementation/test consistency, or breaking PRDs and stories into fixed-format role tasks with sizing…
Skill Claude Code
Bounded owner-scoped repair workflow for selected findings from sdcorejs-review, external human/PR/agent feedback, sdcorejs-ship, supported implementation tracks, simplify output, and product/design/test/documentation/workflow artifacts. Verifies claims, authority, and evidence, repairs within registry-supported…
Skill Claude Code
Read-only review/audit skill for every central-registry track, including Angular, NestJS, Next.js, AI-agent, product, design, documentation, workflow, test, React, Node, fullstack, and general. Reviews code/artifacts, ownership, security, evidence, traceability, and generated/reference validity; emits durable findings…
Skill Claude Code
Final delivery and dependency-update gate. Use for verify-before-done, branch readiness, ready-to-merge, ship/push/release/tag, update dependencies, update package, bump package, npm outdated, audit fix, or consistency checks before handoff. Owns verify-before-done, branch-ready, ship, dependency-update, and…
Skill Claude Code
Behavior-preserving code simplification utility for recently changed or explicitly scoped executable source code. Use to improve clarity, consistency, and maintainability after a green baseline without changing observable behavior, public contracts, strings, prompts, configuration, or protected framework boundaries.…
Skill Claude Code
Spec authoring and approval gate. Use after brainstorming confirms requirements, or when the user asks to write/review/approve/change a spec or design doc. Writes a draft in the matching track directory under .sdcorejs/docs, self-reviews, waits for explicit approval, then persists the approved snapshot in the matching…
Skill Claude Code
Requirement-driven test executor for planning, authoring, running, TDD, UAT, test coverage analysis, authenticated browser testing, and verified UI evidence across existing project stacks. Use for direct test work; route product requirement/traceability coverage without test work to sdcorejs-product, debugging fixes…
Skill Claude Code
Session bootstrap and dispatch for sdcorejs skills. Use at session start, onboarding/help/list-skill requests, or matching skill intent. Routes Q&A, bounded fast-fixes and governed execution with approval, verification and finish gates. Runtime-localized.
Skill Codex needs its repo
Use when changing this repository's skill inventory, skill triggers, dispatch behavior, shared skill contracts, or behavioral evals. Internal to sdcorejs-agent authoring; never install or distribute it to target projects.
Plugin Claude Code
Engineering-focused AI-assisted SDLC for developers and technical teams: clarify, specify, plan, implement through technical tracks, test, review, and ship with explicit approval gates and evidence. Skills are auto-synced from the canonical source tree and localize output at runtime.
Skill Claude Code needs its repo
Approved-plan executor for first-class AI-agent contracts using one declared engine profile and one independent capability profile. Use for confirmed implementation of Responses API or Agents SDK agent architecture, tools, guardrails, approvals, state, evidence, tracing, offline evals, and tests. Under-specified agent…
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: