# Use Linear CLI

> Use when driving Linear through linear-cli for issues, bulk operations, search, or git and PR workflows.

- Skill: `yigitkonur/use-linear-cli` (Agent Skill, multi-file: 21 files)
- Install (CLI): `npx get-skill add yigitkonur/use-linear-cli`
- Raw SKILL.md: https://api.skillmd.com/api/skills/yigitkonur/use-linear-cli/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: yigitkonur (https://skillmd.com/u/yigitkonur)
- Updated: 2026-11-11
- Page: https://skillmd.com/skills/yigitkonur/use-linear-cli

---


# Use linear-cli

Drive [nesszer/linear-cli](https://github.com/nesszer/linear-cli), whose executable is `linear-cli`. This skill targets release **0.3.28**. Check `linear-cli --version`; use the installed command's help when versions differ.

Read [live compatibility](references/live-compatibility.md) before using remote templates, time tracking, or bulk previews.

## Workflow

1. **Discover unfamiliar commands/flags** through the installed command’s `--help`. [command-surface.md](references/command-surface.md) indexes the public tree. Global flag acceptance does not prove handler support.
2. **Verify identity once per session and after profile changes** with `linear-cli u me --output json`. `auth status --validate` can exit 0 without validated credentials; see [setup.md](references/setup.md) for auth precedence.
3. **Resolve scope.** Resolve team keys, assignees, labels, workflow states, project names, and issue identifiers from actual data. Never pick the first team or invent IDs. Respect the user's chosen tool and workspace.
4. **Read before changing.** Fetch current values and the exact target set. For generated bulk sets, save and inspect IDs, count them, and preview the intended changes. Use `--dry-run` where supported; its coverage differs by command. Keep the concrete target set within the user’s authorization.
5. **Execute and reconcile.** Capture successful IDs and per-item results. Re-read changed resources. Partial failures are not transactions: restore a previous value only if it was captured and the follow-up change is authorized.

## Output rules

- Request `--output json` for structured results, and keep stdout and stderr separate. Runtime JSON errors go to **stderr**; parser failures are plain text. Some handlers still emit prose or no JSON.
- Use `--id-only` only on handlers that implement it, such as issue create/update. `i list --id-only` does not produce a plain ID list; extract identifiers from JSON with `jq`.
- Set `LINEAR_CLI_NO_PAGER=true` in agent runs. Use `--yes` or a command's `--force` only for an authorized operation; there is no global `-y` alias.

## Common reads

```bash
linear-cli --version
linear-cli u me --output json
linear-cli t list --output json
linear-cli i list --mine -t ENG --output json --fields identifier,title,state.name
linear-cli i get LIN-123 --output json
linear-cli context --output json
```

`ENG` and `LIN-123` are placeholders to replace with resolved values. `i list` defaults to a limited page and does not filter to open issues. `i get LIN-1 LIN-2` performs concurrent per-ID requests; JSON results may be sorted and partial failures still exit nonzero.

## Route to the task

Read only the relevant reference. Bash recipes use Bash syntax; the checklist recipes support macOS's Bash 3.2.

| Task | Reference |
|---|---|
| Checklist, CSV, JSON, template, or parent/child bulk creation | [creating-many-issues.md](references/recipes/creating-many-issues.md) |
| Start work, create a GitHub PR, branch context, jj, stop work | [git-and-pr-loop.md](references/recipes/git-and-pr-loop.md) |
| Triage queue, comments, screenshots, attachments | [triage-and-comments.md](references/recipes/triage-and-comments.md) |
| Issue create/update, JSON input, transitions, move/transfer, archive/delete | [lifecycle.md](references/issues/lifecycle.md) |
| Search/list filters, saved views, favorites | [search-and-filter.md](references/issues/search-and-filter.md) |
| Relations, labels, statuses, local and remote templates | [relations-and-labels.md](references/issues/relations-and-labels.md) |
| Projects, updates, cycles, sprints, milestones, roadmaps, initiatives | [projects-and-cycles.md](references/planning/projects-and-cycles.md) |
| Teams, users, views, favorites | [teams-and-org.md](references/planning/teams-and-org.md) |
| Import/export formats, duplicate creation, preview and partial-failure limits | [import-export.md](references/data/import-export.md) |
| Attachment CRUD, authenticated upload downloads, URL extraction | [attachments-and-uploads.md](references/data/attachments-and-uploads.md) |
| Watch events, temporary webhooks, notifications, metrics, history, time | [eventing-and-tracking.md](references/eventing-and-tracking.md) |
| Raw GraphQL, pagination, documents | [advanced.md](references/advanced.md) |
| Installation, auth, profiles, config paths, diagnostics, completions | [setup.md](references/setup.md) |
| Output handling, stdin, pagination, retries, safe chaining | [output-and-scripting.md](references/output-and-scripting.md) |
| JSON shapes and command-specific limitations | [json-shapes.md](references/json-shapes.md) |
| Auth errors, rate limits, partial results, missing commands, cache | [troubleshooting.md](references/troubleshooting.md) |
| Full command discovery and local verification | [command-surface.md](references/command-surface.md) |