orbit-knowledge-graph

Index and query a LOCAL checkout of a repository offline with the Orbit local CLI (the `orbit` binary, run directly or via `glab orbit local`). It builds a DuckDB property graph from the working tree and you query it with read-only SQL. Use when the request targets the current checkout, working tree, or a branch that is not pushed/indexed remotely, or is explicitly offline/local: index this repo locally, who calls X in my checkout, list definitions in a file, generate a repo map of a local checkout, run SQL over the local code graph, or serve the local graph over MCP. For queries against already-indexed production data in GitLab (a project such as gitlab-org/gitlab, cross-project blast radius, contributor or merge-request aggregation) use the `orbit` skill; for single-entity GitLab lookups or write operations use `glab`.

AST-based code search and rewrite via tree-sitter patterns. Use instead of Grep/Edit when a change spans many sites or must respect code structure: structural matching, multi-file batch rewrites, or context-aware queries (e.g. "unwrap inside impl blocks"). Not for a single-file or single-symbol rename — use Edit for that.

Registry of related repositories (Siphon, Rails, and other consumed repos) with their git URLs and local checkout paths. Use when you need to find where a related repo is checked out locally, get its git URL, or clone it if it is missing.

Query the GitLab Knowledge Graph (Orbit) via `glab orbit remote` CLI subcommands or run a local copy with `glab orbit local`. Use for code-structure questions (who calls this function, where is this symbol defined), cross-project dependency and blast-radius analysis, merge-request and contributor queries that require relationship traversal or aggregation, repository map / repo-map generation, and any question spanning relationships, cross-entity joins, or multi-entity aggregation across GitLab entities (projects, users, MRs, issues, pipelines, files, definitions, vulnerabilities). Do not use for single-entity GitLab lookups or write operations that `glab` handles directly (e.g. `glab mr view`, `glab mr create`).

Remove and tighten unnecessary LLM-generated comments — including narration disguised as "why", restated constraints, and why-comments a named symbol or the code structure already conveys. Use while editing, and as the final narration-comment pass before opening an MR or pushing, to strip the LLM narration the model left behind.

Audit and update documentation after code changes. Use when architecture, APIs, or behavior changed and docs may have drifted.

Investigate query evaluation failures in the Knowledge Graph synthetic data pipeline. Use when queries fail or return unexpected results after running the evaluate binary.

Profile GKG queries against ClickHouse with the query-profiler CLI. For optimizing query performance, comparing query plans, investigating slow queries, or checking ClickHouse resource usage.

Investigate the history, usage, and liveness of code using search and git blame/log. Use when determining if code is dead, understanding why something exists, finding all callers before refactoring, or deciding whether something is safe to remove. Also useful for answering "who added this and why" or "is anything still using this".

GitLab Pajamas Design System expert for building UIs with Pajamas components and patterns. Use when: (1) implementing UI that should follow GitLab's Pajamas design system, (2) selecting or configuring Pajamas/GlComponent components (GlButton, GlAlert, GlModal, etc.), (3) translating Figma designs into Pajamas-compliant code, (4) questions about Pajamas component usage, variants, categories, or accessibility, (5) building GitLab-style interfaces, or (6) the user mentions "Pajamas", "GitLab UI", "Gl components", or "design system" in a GitLab context. Works hand-in-hand with the implement-design skill and Figma MCP tools.

Trace and document how data transforms through a multi-step pipeline or function chain, showing intermediate state at each step with concrete example values. Use when explaining a data pipeline or complicated codepaths, tracing how a value changes across function calls, answering questions like "how does X get to Y", or producing a step-by-step dataflow walkthrough for a code review or design doc.