> For the complete documentation index, see [llms.txt](https://developer.celigo.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://developer.celigo.com/cli/using-the-cli/skills.md). # AI assistant skills Installing the Celigo CLI also installs a set of domain-specific **skills** that AI coding assistants can load on demand. When you ask Claude Code, Cursor, or Windsurf to "build a Shopify → NetSuite flow" or "debug this stuck job," the assistant pulls in the relevant skill files, reads the workflow guidance and field schemas inside, and gives grounded answers — instead of guessing about integrator.io. ## Supported assistants The CLI's `postinstall` script symlinks every shipped skill into your assistant's skills directory. Three hosts are wired up today: | Assistant | Global install path | Local (per-project) path | | ----------- | ----------------------------- | ----------------------------- | | Claude Code | `~/.claude/skills/` | `/.claude/skills/` | | Cursor | `~/.cursor/skills/` | `/.cursor/skills/` | | Windsurf | `~/.codeium/windsurf/skills/` | `/.windsurf/skills/` | Only directories that already exist get populated — if you don't have Cursor installed, nothing gets written to `~/.cursor/`. Skills appear prefixed `npm-celigo-cli-` so you can tell Celigo skills apart from your own. The `CLAUDE_CONFIG_DIR` env var is honored on the Claude Code path. On Windows, symlinks are created as directory junctions (no elevated privileges required). ## How they work Each skill directory contains: * **`SKILL.md`** — a workflow guide the assistant reads when the skill triggers. Describes the task scope, the build order, common pitfalls, and concrete examples. * **`references/`** — OpenAPI-derived field schemas (YAML) for the resources the skill touches, so the assistant writes JSON that actually validates against integrator.io. When you ask a relevant question, the assistant selects one or more skills based on the task, loads the `SKILL.md`, and uses the schemas to produce correct configuration. For example: "help me set up an HTTP export with a delta cursor" pulls in `configuring-exports` + `writing-handlebars`. ## What ships **Orientation** * `getting-started` — core concepts, build order, and which skill to use for each task. Start here when a task is unclear or you're new to Celigo. * `troubleshooting-flows` — diagnose flow failures: total failures, partial errors, stuck jobs, empty runs, performance issues. * `managing-users` — invite users, update permissions, configure access levels, enforce MFA/SSO, disable accounts. **Build** * `building-flows` — pipelines that move data from source to destination on a schedule or in response to events. Covers scheduling, chaining, error management, and abstract/instance templating. * `building-apis` — custom HTTP endpoints that let external systems push or query data synchronously through Celigo. * `building-tools` — reusable building blocks that encapsulate lookups, imports, transforms, and branching behind input/output contracts. * `building-mcp-servers` — MCP endpoints that expose Tools and builder-mode APIs to external AI agents and MCP clients. * `building-b2b` — B2B Manager EDI integrations: trading partner onboarding, EDI profiles, file definitions, X12/EDIFACT flow patterns. **Configure** * `configuring-connections` — credentials and configuration objects that authenticate to external systems. Auth methods, OAuth, iClients. * `configuring-exports` — the data source step: choose adaptor type, set up delta/incremental syncs, webhooks, file transfers, lookups. * `configuring-imports` — the destination step: choose adaptor type, mappings, lookups, upsert logic, AI agent and file imports. * `configuring-lookup-caches` — in-memory key-value stores for fast lookups, dedup, cross-reference resolution, state tracking. * `configuring-ai-agents` — LLM-powered flow steps that classify, extract, validate, or generate data. OpenAI, Gemini, BYOK. * `configuring-guardrails` — safety and compliance checks: PII detection, content moderation, AI-based evaluation rules. * `configuring-filters` — filter rules on exports, imports, and flow branches that control which records continue. **Write** * `writing-mappings` — field mappings and transforms: Mapper 2.0 (imports), Transformation 2.0 (exports), lookups, response mapping, Mapper 1.0. * `writing-scripts` — Celigo JavaScript hook scripts: `preSavePage`, `preMap`, `postMap`, `postSubmit`, `postResponseMap`, filter, transform, branching, `handleRequest`. * `writing-handlebars` — Handlebars template expressions for dynamic values in mappings, HTTP bodies, SQL queries, URIs, filters. * `writing-sql` — SQL for RDBMS exports and imports across Snowflake, Postgres, MySQL, SQL Server, Oracle, BigQuery, Redshift. ## Seeing what's installed ```bash # Claude Code ls ~/.claude/skills/ | grep celigo # Cursor ls ~/.cursor/skills/ | grep celigo # Windsurf ls ~/.codeium/windsurf/skills/ | grep celigo ``` Each entry is a symlink pointing back into the CLI package's `skills/` directory. Upgrading the CLI (`npm install -g @celigo/celigo-cli@latest`) updates the skills atomically — the symlinks keep pointing at the package, and the package content changes. ## Opting out There's no env-var flag to suppress skill installation. If you don't want them: * **Skip the postinstall step** — `npm install -g @celigo/celigo-cli --ignore-scripts`. You get the binary without the symlinks. * **Remove after install** — delete the `npm-celigo-cli-*` entries from whichever skills directory you prefer. The symlinks will come back on the next upgrade unless you also use `--ignore-scripts`. ## Schema source of truth Schemas bundled under each skill's `references/schemas/` originate from Celigo's internal `ai` skills repository, the source of truth for skills, which tracks the authoritative OpenAPI specs published to the API reference. The CLI fetches them from a pinned `celigo/ai` ref at build time and bundles them into the published package — so whatever your AI assistant reads in a skill reflects the current API. --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter: ``` GET https://developer.celigo.com/cli/using-the-cli/skills.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.