jpeetz/agent-skills

linear-project-management

Use when driving Linear projects and issues from an agent.

Quelltext ansehen

Verwendung

  1. Installationsbefehl kopieren und das Quellpaket vor der Installation in der eigenen Agentenumgebung prüfen.
  2. Benötigte Werkzeuge, Zugangsdaten und Hosts im Original prüfen.
  3. Eingabe und gewünschtes Ergebnis beschreiben und die Ausgabe mit den Quellen abgleichen.
npx skills add https://github.com/jpeetz/agent-skills --skill linear-project-management

Vor dem Start

  • Die Installation umfasst weder Modellabo noch API-Kontingent oder Dateizugriffsrechte.
  • Sterne und Installationen zeigen Aktivität, keine unabhängige Qualitäts- oder Sicherheitsprüfung.
Originales Skill-Dokument

Aus dem Quell-Repository gerendert; Überschriften, Beispiele, Code, Tabellen, Links und Bilder bleiben erhalten.

Linear Project Management

Drive Linear (linear.app) from an agent: create and update issues, projects, and teams; route work with a label taxonomy; and keep workflows aligned with code changes. It is designed for agents using the official Linear MCP server for simple operations and the Linear GraphQL API for anything more advanced.

Two ground rules save most of the pain: always discover before you create (Linear accumulates duplicates fast) and know the difference between `description` and `content` on a project. This skill encodes both.

When to Use

Use this skill when you want to:

  • Create or update a Linear issue, project, or initiative.
  • Check for an existing issue/project before making a duplicate.
  • Classify and label issues consistently across a team.
  • Move issues through workflow states, or bulk-sync a sprint to Done.
  • Attach a resource link or milestone to a project.

Don't use for: non-Linear ticketing, or when the data model difference (description vs content) does not matter — that is exactly when you should use it, but if you are only reading one status value, a quick GraphQL query is enough.

Prerequisites

  • Linear access token (starts lin_api_... ) exposed as the environment

variable LINEAR_ACCESS_TOKEN. Generate at Linear → Settings → Security & access → Personal API keys. Never hardcode it in files or commits.

  • Official Linear MCP server (recommended for simple ops) wired with the

same token:

json
  {"mcpServers": {"linear": {"command": "npx", "args": ["mcp-remote", "https://mcp.linear.app/sse"], "env": {"LINEAR_API_KEY": "{token}"}}}}

Prefer mcp.linear.app; avoid deprecated community servers.

  • Node.js >= 20 only if you use the @linear/sdk/scripts path (optional).

How to Run

Simple operation via MCP:

Create a high priority bug titled "Fix auth timeout" in the ENG team.

Discovery before create (mandatory):

Search Linear for any existing project matching "checkout-redesign" before creating one.

GraphQL for advanced/authoritative access — run the query through your agent's GraphQL tool with the Linear endpoint (https://api.linear.app/graphql) and Authorization: {LINEAR_ACCESS_TOKEN} header.

Quick Reference

OperationApproach
Create/update issueMCP, or GraphQL issueCreate/issueUpdate
Create projectGraphQL projectCreate (set both description + content)
Find duplicates firstGraphQL issues/projects search by title/keyword
List statusesGraphQL projectStatuses { nodes { id name } }
MilestoneprojectMilestoneCreate
Resource linkentityExternalLinkCreate
Bulk state changeissueUpdate loop, or @linear/sdk script

Procedure

  1. Discover before create. Before any new issue/project, search Linear by

title keywords and fail fast if a match exists, then propose an update instead. Completion criterion: a search ran and returned either an existing match (handled) or a confirmed empty result.

  1. Pick the write path. Simple = MCP. Authoritative/multi-field = GraphQL.

Completion criterion: one path is chosen and the payload is mapped to the chosen shape's fields.

  1. Set both content fields on projects. description (≤255 chars, shows in

list views) and content (unlimited, shows in the main detail panel). If you set only one, half the interface is blank. Stop and fill both. Completion criterion: both fields populated for projects/initiatives.

  1. Assign labels by taxonomy. Choose exactly one Type (feature, bug,

refactor, chore, spike); 1-2 Domain tags (e.g. security, backend, frontend, testing, infrastructure); 0-2 Scope tags (blocked, breaking-change, tech-debt, needs-split, good-first-issue). Completion criterion: the issue carries one Type and at most 2+1 domain/scope.

  1. Resolve states against workspace statuses. Status UUIDs are

workspace-specific — query projectStatuses first, never hardcode an ID. Completion criterion: status IDs were fetched, not assumed.

  1. If bulk-syncing a sprint: batch issueUpdate to Done and update the

project, then verify a sample. Completion criterion: at least 2 issues verified as Done after the run.

Pitfalls

  • Duplicate issues. The number-one failure. Always search before create.
  • `description` vs `content`. Set both; one without the other renders blank.
  • Hardcoded status UUIDs. They are workspace-specific. Always query.
  • Hardcoded API token. Keep it in LINEAR_ACCESS_TOKEN, never in the skill,

an issue body, or a repo.

  • Unchecked codebase scope. Issue text describing "missing" features may

already be implemented — verify the codebase before re-implementing.

Verification

  • [ ] Creating an issue returns its ID and the matching label Type.
  • [ ] A project create carries non-empty description AND content.
  • [ ] A "find duplicates" query returns the pre-existing issue when it exists.
  • [ ] Status IDs come from projectStatuses, not a constant.
  • [ ] No call in the transcript echoes or writes the raw token.
aus demselben Repository

Weitere Skills

Alle Skills
jpeetz
Community

accessibility-compliance-audit

Comprehensive web accessibility (a11y) audit and compliance skill. Audits HTML, React, Vue, and Angular codebases against WCAG 2.2 AA standards. Activates when users ask for "accessibility audit", "a11y check", "WCAG compliance", "ADA compliance", "screen reader testing", "keyboard navigation audit", "color contrast check", "accessibility fix", or "make this accessible". Detects violations, explains impact, and generates fix-ready code. Covers semantic HTML, ARIA authoring, focus management, col

Installationen
1
GitHub Stars
14
Aktualisiert
21. Okt.
jpeetz
Community

agent-n-video-pipeline

Use this skill when producing a fully-local "Agent N presents a message" video on the 16GB Apple-Silicon M5 executor: Kokoro mature TTS → LTX-2.3 MLX A2V talking render with native lip sync → karaoke captions → trim → QA → X caption. Replaces the old cloud Herald pipeline (grok-imagine + volcengine) with zero cloud cost. Activates for Herald release videos, Agent N announcements, and newsletter video messages.

Installationen
1
GitHub Stars
14
Aktualisiert
21. Okt.
jpeetz
Community

anti-over-engineering

Use when AI agents over-engineer code beyond user requests.

Installationen
1
GitHub Stars
14
Aktualisiert
21. Okt.
jpeetz
Community

api-and-interface-design

Use this skill when designing APIs, module boundaries, or any public interface. Create stable, well-documented interfaces that are hard to misuse — REST endpoints, GraphQL schemas, module contracts, component props. Covers contract-first design, error semantics, validation boundaries, idempotency, and consistent naming.

Installationen
1
GitHub Stars
14
Aktualisiert
21. Okt.