/api
GitLab REST and GraphQL API access via glab. Use when making API requests, running GraphQL queries or mutations, querying project data, automating GitLab operations, or looking up GitLab feature, API, or CI/CD documentation.
$ npx -y skills add bendrucker/claude --skill api --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
- Fires itselfAuto-invocation. Claude auto-loads it when your prompt matches the work.Auto-invocation is when the right skill fires by itself at the right moment, driven by a FLOW.md router and a hook, instead of you invoking it by name. It is the difference between a skill being installed and a skill actually getting used.Read the full definition →
- You can call itInvoke it directly when you want it.
- Slash command
/api
Context preview
The summary Claude sees to decide when to auto-load this skill.
GitLab REST and GraphQL API access via glab. Use when making API requests, running GraphQL queries or mutations, querying project data, automating GitLab operations, or looking up GitLab feature, API, or CI/CD documentation.
SKILL.md
api.SKILL.mdname: gitlab:api
description: GitLab REST and GraphQL API access via glab. Use when making API requests, running GraphQL queries or mutations, querying project data, automating GitLab operations, or looking up GitLab feature, API, or CI/CD documentation.
GitLab API
REST and GraphQL API access via `glab api`.
Placeholders
Auto-resolve to current project values:
- `:fullpath` - Full project path (e.g., `group/project`)
- `:id` - Project ID
- `:branch` - Current branch
- `:user` / `:username` - Current user
glab api projects/:fullpath/merge_requests
REST
glab api projects/:id/issues # GET
glab api projects/:id/issues -X POST -f title="..." # POST with field
glab api projects/:id/issues --paginate # All pages
No `--jq`/`-q`
Output filtering flags belong to `gh api`. Pipe to `jq` instead: `glab api <endpoint> | jq '<filter>'`. There is also no `--json` flag anywhere in `glab`. Use `--output json`.
Error Bodies Land on Stdout
A 404/4xx response is a JSON error body (`{"message":"404 Not found"}`), so `glab api ... | jq '.field'` turns the real HTTP error into misleading `null` output or a jq type error (`Cannot index string`, `Cannot iterate over object`). When a pipeline misbehaves, rerun the `glab api` call bare and read the body before touching the jq filter.
Pagination Pitfall
`--paginate` concatenates JSON arrays across pages as `][`, producing invalid JSON (e.g., `[{...}][{...}]`). Fix by replacing `][` with `,` before parsing: `.replace(/\]\s*\[/g, ",")`.
Nested Fields
`-f` and `--raw-field` silently drop bracket-nested keys like `position[base_sha]=...`. For nested objects, write JSON to a file and use `--input`:
echo '{"position":{"base_sha":"abc","head_sha":"def","old_path":"file.ts","new_path":"file.ts","position_type":"text","new_line":10}}' > /tmp/payload.json
glab api projects/:id/merge_requests/:iid/discussions -X POST -H "Content-Type: application/json" --input /tmp/payload.jsonUploads
An image or video for an MR or issue body goes through project uploads, which take multipart form data. `--form` sends the file. `-F` reads the file and embeds its bytes in a JSON body field, which the endpoint rejects with HTTP 400.
glab api projects/:id/uploads --form "file=@./shot.png" | jq -r .markdown
The response carries `url` (a project-relative `/uploads/<hash>/shot.png`) and `markdown` (``). Paste the markdown into the description or note. The relative URL renders anywhere inside that project. `glab mr` and `glab issue` have no attach flag. The upload runs before the create or update.
GraphQL
glab api graphql -f query='{ currentUser { username } }'For pagination, accept `$endCursor` variable and fetch `pageInfo { hasNextPage, endCursor }`.
Queries With Variables: Use a Quoted Heredoc
Escaping `$` in an inline `-f query='mutation($var...)'` corrupts the query (GitLab returns `Expected VAR_SIGN, actual: UNKNOWN_CHAR`). A quoted heredoc needs no escaping:
glab api graphql -f query="$(cat <<'GQL'
mutation($projectPath: ID!, $iid: String!, $userId: UserID!) {
mergeRequestReviewerRereview(input: { projectPath: $projectPath, iid: $iid, userId: $userId }) { errors }
}
GQL
)" -f projectPath=group/project -f iid=123 -f userId="gid://gitlab/User/42"Pass variables with `-f` (strings), not `-F`: `-F` coerces numeric-looking values to int, and `iid` is typed `String!`. `projectPath` is typed `ID!`, not `String!`.
Hallucinated Mutations
Do not guess mutation names from GitHub's schema. `mergeRequestSetAutoMerge` and `mergeRequestRequestReview` are not in GitLab's schema. Auto-merge is REST-only (`glab mr merge --auto-merge`, or the `gitlab:merge-request` skill's `merge.ts` for merge trains). Re-requesting review is `mergeRequestReviewerRereview`. Requesting changes is `mergeRequestRequestChanges`. Trial-executing a guessed mutation runs it for real if it exists, so use documented forms only.
Output
- `--output json` (default) - Pretty-printed JSON
- `--output ndjson` - Newline-delimited, works with `jq` streaming
Documentation
For `glab` command reference, use `glab <command> --help` instead of web docs (the `/cli/` pages duplicate CLI help output). For everything else, [GitLab docs](https://docs.gitlab.com/) paths:
| Need | Path | |------|------| | REST endpoints | `/api/` | | MR API fields | `/api/merge_requests/` | | GraphQL schema | `/api/graphql/reference/` | | CI/CD YAML syntax | `/ci/yaml/` | | Predefined CI variables | `/ci/variables/` | | Feature concepts | `/topics/` | | UI configuration | `/user/` | | Self-managed admin | `/administration/` |
Read more
name: gitlab:api description: GitLab REST and GraphQL API access via glab. Use when making API requests, running GraphQL queries or mutations, querying project data, automating GitLab operations, or looking up GitLab feature, API, or CI/CD documentation.
GitLab API
REST and GraphQL API access via `glab api`.
Placeholders
Auto-resolve to current project values:
- `:fullpath` - Full project path (e.g., `group/project`)
- `:id` - Project ID
- `:branch` - Current branch
- `:user` / `:username` - Current user
glab api projects/:fullpath/merge_requests
REST
glab api projects/:id/issues # GET glab api projects/:id/issues -X POST -f title="..." # POST with field glab api projects/:id/issues --paginate # All pages
No `--jq`/`-q`
Output filtering flags belong to `gh api`. Pipe to `jq` instead: `glab api <endpoint> | jq '<filter>'`. There is also no `--json` flag anywhere in `glab`. Use `--output json`.
Error Bodies Land on Stdout
A 404/4xx response is a JSON error body (`{"message":"404 Not found"}`), so `glab api ... | jq '.field'` turns the real HTTP error into misleading `null` output or a jq type error (`Cannot index string`, `Cannot iterate over object`). When a pipeline misbehaves, rerun the `glab api` call bare and read the body before touching the jq filter.
Pagination Pitfall
`--paginate` concatenates JSON arrays across pages as `][`, producing invalid JSON (e.g., `[{...}][{...}]`). Fix by replacing `][` with `,` before parsing: `.replace(/\]\s*\[/g, ",")`.
Nested Fields
`-f` and `--raw-field` silently drop bracket-nested keys like `position[base_sha]=...`. For nested objects, write JSON to a file and use `--input`:
echo '{"position":{"base_sha":"abc","head_sha":"def","old_path":"file.ts","new_path":"file.ts","position_type":"text","new_line":10}}' > /tmp/payload.json
glab api projects/:id/merge_requests/:iid/discussions -X POST -H "Content-Type: application/json" --input /tmp/payload.jsonUploads
An image or video for an MR or issue body goes through project uploads, which take multipart form data. `--form` sends the file. `-F` reads the file and embeds its bytes in a JSON body field, which the endpoint rejects with HTTP 400.
glab api projects/:id/uploads --form "file=@./shot.png" | jq -r .markdown
The response carries `url` (a project-relative `/uploads/<hash>/shot.png`) and `markdown` (``). Paste the markdown into the description or note. The relative URL renders anywhere inside that project. `glab mr` and `glab issue` have no attach flag. The upload runs before the create or update.
GraphQL
glab api graphql -f query='{ currentUser { username } }'For pagination, accept `$endCursor` variable and fetch `pageInfo { hasNextPage, endCursor }`.
Queries With Variables: Use a Quoted Heredoc
Escaping `$` in an inline `-f query='mutation($var...)'` corrupts the query (GitLab returns `Expected VAR_SIGN, actual: UNKNOWN_CHAR`). A quoted heredoc needs no escaping:
glab api graphql -f query="$(cat <<'GQL'
mutation($projectPath: ID!, $iid: String!, $userId: UserID!) {
mergeRequestReviewerRereview(input: { projectPath: $projectPath, iid: $iid, userId: $userId }) { errors }
}
GQL
)" -f projectPath=group/project -f iid=123 -f userId="gid://gitlab/User/42"Pass variables with `-f` (strings), not `-F`: `-F` coerces numeric-looking values to int, and `iid` is typed `String!`. `projectPath` is typed `ID!`, not `String!`.
Hallucinated Mutations
Do not guess mutation names from GitHub's schema. `mergeRequestSetAutoMerge` and `mergeRequestRequestReview` are not in GitLab's schema. Auto-merge is REST-only (`glab mr merge --auto-merge`, or the `gitlab:merge-request` skill's `merge.ts` for merge trains). Re-requesting review is `mergeRequestReviewerRereview`. Requesting changes is `mergeRequestRequestChanges`. Trial-executing a guessed mutation runs it for real if it exists, so use documented forms only.
Output
- `--output json` (default) - Pretty-printed JSON
- `--output ndjson` - Newline-delimited, works with `jq` streaming
Documentation
For `glab` command reference, use `glab <command> --help` instead of web docs (the `/cli/` pages duplicate CLI help output). For everything else, [GitLab docs](https://docs.gitlab.com/) paths:
| Need | Path | |------|------| | REST endpoints | `/api/` | | MR API fields | `/api/merge_requests/` | | GraphQL schema | `/api/graphql/reference/` | | CI/CD YAML syntax | `/ci/yaml/` | | Predefined CI variables | `/ci/variables/` | | Feature concepts | `/topics/` | | UI configuration | `/user/` | | Self-managed admin | `/administration/` |
My personal plugin marketplace for Claude Code, Anthropic's AI coding assistant.
Repo: bendrucker/claude
Other skills on bendrucker-claude.
cleye
Type-safe CLI argument parsing with cleye, the standard parser for this repo's Bun scripts. Use when writing or editing any script that takes arguments (flags,…
activity
Report real device usage from ActivityWatch. Covers per-app time, window titles, and active vs idle spans. Use when asked "what apps did I use", "how long was…
history
Report shell history from atuin's local capture. Covers what commands ran, when, where, and how they exited. Use when asked "what commands did I run", "what…
bun
Bun runtime patterns. Use when running bun commands, working with package.json/bun.lock, writing TypeScript scripts under Bun, or developing Claude Code…
agent-team
Orchestrating Claude Code agent teams. Use when creating teams, spawning teammates, assigning tasks, configuring teammate modes, or setting up team quality…

