Skip to content
Back to skills

Linear Sop

ASecurity

Linear ticket management best practices. Use when creating issues, updating status, or attaching evidence. Provides evidence templates for dev/staging/done phases.

  • 404 stars
  • 0 votes
  • 0 copies
  • 1 view
  • Added February 7, 2026
developmenttypescriptbashtestinggitapidocumentation

Works with

  • claude code
  • api
  • mcp

Security analysis

A100/100

Pro scans all 2 files and shows the line behind each finding

Scanned February 12, 2026

npx -y skills add bybren-llc/safe-agentic-workflow --skill linear-sop --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Linear Sop?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Linear Sop
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/bybren-llc-linear-sop/badge)](https://www.skillsdirectory.com/skills/bybren-llc-linear-sop)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: linear-sop
description: Linear ticket management best practices. Use when creating issues, updating status, or attaching evidence. Provides evidence templates for dev/staging/done phases.
---

# Linear SOP Skill

## Purpose

Guide consistent Linear ticket management. Provides evidence templates for the mandatory dev/staging/UAT evidence policy.

## When This Skill Applies

Invoke this skill when:

- Creating new Linear issues
- Updating ticket status
- Attaching evidence to tickets
- Parsing acceptance criteria
- Working with UUIDs and issue IDs

## Linear MCP Tools

### Reading Issues

```text
# Get issue by identifier
mcp__linear-mcp__get_issue({ id: "{{TICKET_PREFIX}}-459" })

# List issues with filters
mcp__linear-mcp__list_issues({
  team: "{{PROJECT_TEAM_NAME}}",
  state: "In Progress",
  assignee: "me",
})
```

### Creating Issues

```text
mcp__linear-mcp__create_issue({
  title: "feat(scope): description",
  team: "{{PROJECT_TEAM_NAME}}",
  description: "## Summary\n\n...",
  labels: ["feature", "sprint-1"],
  parentId: "parent-uuid",  // Optional - for sub-issues
})
```

### Updating Issues

```text
mcp__linear-mcp__update_issue({
  id: "{{TICKET_PREFIX}}-459",
  state: "Done",
})
```

### Adding Comments

```text
mcp__linear-mcp__create_comment({
  issueId: "{{TICKET_PREFIX}}-459",
  body: "**Dev Evidence**\n\n...",
})
```

## Evidence Policy (MUST)

Every issue requires evidence at each phase:

| Phase       | Required? | Content                 |
| ----------- | --------- | ----------------------- |
| **Dev**     | MUST      | Implementation proof    |
| **Staging** | MUST      | UAT validation (or N/A) |
| **Done**    | MUST      | Final verification      |

## Evidence Templates

### Dev Evidence Template

```markdown
**Dev Evidence**

**PR**: https://github.com/{{ORG_NAME}}/{{REPO_NAME}}/pull/XXX
**Commit**: [short-hash]
**Branch**: {{TICKET_PREFIX}}-XXX-description

**Implementation:**

- [x] Feature implemented
- [x] Tests passing
- [x] Lint passing

**Verification:**

\`\`\`bash
yarn ci:validate

# Output: All checks passed

\`\`\`
```

### Staging/UAT Evidence Template

```markdown
**Staging Evidence**

**Environment**: Pop OS dev server
**URL**: http://pop-os:3000

**Validation Steps:**

1. Deployed to staging: [timestamp]
2. Smoke test passed: [yes/no]
3. Feature verified: [description]

**UAT Status:** [Passed/Pending/N/A]

If N/A, reason: [e.g., "Dev tooling only - no user-facing changes"]
```

### Done Evidence Template

```markdown
**Done Evidence**

**PR Merged**: https://github.com/{{ORG_NAME}}/{{REPO_NAME}}/pull/XXX
**Merge Commit**: [hash]

**Final Checklist:**

- [x] All acceptance criteria met
- [x] Documentation updated (if applicable)
- [x] No regressions detected
```

## Acceptance Criteria Parsing

When reading issue descriptions, extract ACs:

```markdown
## Acceptance Criteria

- [ ] User can perform action X
- [ ] System responds with Y
- [ ] Error handling for Z
```

Convert to testable checklist:

```typescript
const acceptanceCriteria = [
  { criterion: "User can perform action X", verified: false },
  { criterion: "System responds with Y", verified: false },
  { criterion: "Error handling for Z", verified: false },
];
```

## Status Workflow

```text
Backlog -> Ready -> In Progress -> Testing -> Ready for Review -> Done
```

### Status Update Guidelines

| From             | To               | When                     |
| ---------------- | ---------------- | ------------------------ |
| Backlog          | Ready            | Sprint planning          |
| Ready            | In Progress      | Work starts              |
| In Progress      | Testing          | PR created               |
| Testing          | Ready for Review | Tests pass, UAT complete |
| Ready for Review | Done             | POPM approval            |

## UUID Handling

Linear uses UUIDs internally. When working with APIs:

```typescript
// Issue identifiers (human-readable)
const issueId = "{{TICKET_PREFIX}}-459";

// UUIDs (API operations)
const uuid = "ef6a5fa0-2b46-417f-8266-dea2d187b10a";

// Get UUID from identifier via MCP tool
// mcp__linear-mcp__get_issue({ id: "{{TICKET_PREFIX}}-459" })
// Returns issue object with .id property containing UUID
```

## Common Operations

### Link PR to Issue

PRs are automatically linked when:

- Branch name contains `{{TICKET_PREFIX}}-XXX`
- PR title contains `[{{TICKET_PREFIX}}-XXX]`

### Create Sub-Issue

```text
mcp__linear-mcp__create_issue({
  title: "Sub-task description",
  team: "{{PROJECT_TEAM_NAME}}",
  parentId: "parent-issue-uuid",
})
```

### Query by Label

```text
mcp__linear-mcp__list_issues({
  label: "sprint-1",
  team: "{{PROJECT_TEAM_NAME}}",
})
```

## Authoritative References

- **Agent Workflow SOP**: `docs/sop/AGENT_WORKFLOW_SOP.md`
- **Linear MCP Docs**: Built into Claude Code
- **CONTRIBUTING.md**: Workflow documentation

Files in this skill

  • README.md1.5 KB
  • SKILL.md4.8 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.