Name: taskNebula
Author: neuraparse

Imported from neuraparse/taskNebula (AGENTS.md). Install upstream with npx skills add neuraparse/taskNebula. Copyright stays with the author.

Agent Instructions

The canonical guide for AI agents and assistants working in this repository is CLAUDE.md — it is the single source of truth for the tech stack, build/test commands, project conventions, the git & release workflow, and architecture.

Tools that follow the AGENTS.md convention should read CLAUDE.md as the full instruction set.

Operator checkouts may also contain an ignored AGENTS.local.md. If present, read it after CLAUDE.md; it contains private deployment context and must never be committed, quoted in public issues/PRs, or copied into tracked source.

i18n is mandatory. The repository contains 30 locale catalogs with device auto-detection; catalog/ICU parity is automated while legacy linguistic review continues. Never hardcode user-facing strings — route every one through next-intl and add the key to all 30 catalogs in apps/web/messages/*.json (key parity: pnpm i18n:check). See .claude/rules/frontend.md and .cursor/rules/i18n.mdc. Applies to all assistants (Claude/Cursor/Codex/Copilot).

Per-package guides (each CLAUDE.md has a sibling AGENTS.md pointer):

apps/web/CLAUDE.md — Next.js app: route auth idiom, validation, i18n, design system
packages/db/CLAUDE.md — Drizzle schema & hand-written migration conventions
packages/mcp-server/CLAUDE.md — MCP tools, API-key boundary, publish gates
services/hocuspocus/CLAUDE.md — Yjs realtime collab server

Install taskNebula and it becomes one more candidate the router can pick - when it fits.

A listing is a routing target with a manifest: what it does, which domains it covers, what it costs and who publishes it. Once installed it sits beside your own models and tools, is scored like any other candidate for each request, and shows up in the trace when it wins. Ratings come from workspaces that installed it, one per account.

{ "ocm": "1", "id": "neuraparse-tasknebula-tasknebula", "kind": "skill", "name": "taskNebula", "description": "The canonical guide for AI agents and assistants working in this repository is **[CLAUDE.md](CLAUDE.md)** — it is the single source of truth for the tech stack, build/test commands, project conventions, the git & release workflow, and architecture.", "publisher": "neuraparse", "version": "1.0.0", "capabilities": { "domains": [ "coding" ], "tags": [ "skill-md", "github-agents" ], "languages": [ "en" ] }, "quality_prior": 0.6, "examples": [ "The canonical guide for AI agents and assistants working in this repository is **[CLAUDE.md](CLAUDE.md)** — it is the single source of truth for the tech stack, build/test commands, project conventions, the git & release workflow, and architecture." ], "primary": false, "metadata": { "source": { "provider": "github-agents", "repository": "https://github.com/neuraparse/taskNebula", "path": "AGENTS.md", "ref": "99b85fa39416ed3c297800e7722b58e37426fb4c", "url": "https://github.com/neuraparse/taskNebula/blob/99b85fa39416ed3c297800e7722b58e37426fb4c/AGENTS.md", "key": "neuraparse/taskNebula/AGENTS.md" } }, "instructions": "# Agent Instructions\n\nThe canonical guide for AI agents and assistants working in this repository\nis **[CLAUDE.md](CLAUDE.md)** — it is the single source of truth for the\ntech stack, build/test commands, project conventions, the git & release\nworkflow, and architecture.\n\nTools that follow the `AGENTS.md` convention should read `CLAUDE.md` as the\nfull instruction set.\n\nOperator checkouts may also contain an ignored `AGENTS.local.md`. If present,\nread it after `CLAUDE.md`; it contains private deployment context and must never\nbe committed, quoted in public issues/PRs, or copied into tracked sour", "cost": { "context_tokens": 391 } }

taskNebula

About

Agent Instructions

Use it

Install taskNebula and it becomes one more candidate the router can pick - when it fits.

Manifest

Reviews