N8n:linear Issue

Start work on Linear issue $ARGUMENTS

Prerequisites

This skill depends on external tools. Before proceeding, verify availability:

Required:

• Linear MCP: Must be connected. Without it the skill cannot function at all.
• GitHub CLI (gh): Must be installed and authenticated. Run gh auth status to verify. Used to fetch linked PRs and issues.

Optional (graceful degradation):

• Notion MCP: Needed only if the issue links to Notion docs. If unavailable, note the Notion links in the summary and tell the user to check them manually.
• Loom transcript skill: Needed only if the issue contains Loom videos. If unavailable, note the Loom links in the summary for the user to watch.
• curl: Used to download images. Almost always available; if missing, skip image downloads and note it.

If a required tool is missing, stop and tell the user what needs to be set up before continuing.

Instructions

Follow these steps to gather comprehensive context about the issue:

1. Fetch the Issue and Comments from Linear

Use the Linear MCP tools available in the active harness to fetch the issue details and comments together:

• Fetch the issue by ID to get full details including attachments
• Include relations to see blocking/related/duplicate issues
• Immediately after, fetch all comments for the issue ID

Both calls should be made together in the same step to gather the complete context upfront.

2. Check for Private/Security Issues (MANDATORY — do this before anything else)

After fetching the issue, immediately check its labels:

1. Look at the labels returned with the issue.

2. If any label is n8n-private: a. Run git remote -v (via Bash) to list all configured remotes. b. If any remote URL contains n8n-io/n8n without the -private suffix (i.e. matches the public repo), stop immediately and tell the user:

   This issue is marked n8n-private and must be developed in a clean clone of the private repository.

   One or more of your remotes point to the public n8n-io/n8n repo. Mixed remotes are not allowed — you must work in a separate local clone of n8n-io/n8n-private with no references to the public repo. For the full process, see: https://www.notion.so/n8n/Processing-critical-high-security-bugs-vulnerabilities-in-private-2f45b6e0c94f803da806f472111fb1a5 (opens in a new tab)

   Do not continue with any further steps — return after showing this message.

3. If the label is not present, or all remotes point exclusively to n8n-io/n8n-private, continue normally.

3. Analyze Attachments and Media (MANDATORY)

IMPORTANT: This step is NOT optional. You MUST scan and fetch all visual content from BOTH the issue description AND all comments.

Screenshots/Images (ALWAYS fetch):

1. Scan the issue description AND all comments for ALL image URLs:
   • <img> tags
   • Markdown images ![](url)
   • Raw URLs (github.com/user-attachments, imgur.com, etc.)
2. For EACH image found (in description or comments):
   • Download using curl -sL "url" -o /path/to/image.png (GitHub URLs require following redirects) OR the linear mcp
   • Use the Read tool on the downloaded file to view it
   • Describe what you see in detail
3. Do NOT skip images - they often contain critical context like error messages, UI states, or configuration

Loom Videos (ALWAYS fetch transcript):

1. Scan the issue description AND all comments for Loom URLs (loom.com/share/…)
2. For EACH Loom video found (in description or comments):
   • Use the Loom transcript skill to fetch the FULL transcript
   • Summarize key points, timestamps, and any demonstrated issues
3. Loom videos often contain crucial reproduction steps and context that text alone cannot convey

4. Fetch Related Context

Related Linear Issues:

• Use the Linear MCP issue-fetching tool for any issues mentioned in relations (blocking, blocked by, related, duplicates)
• Summarize how they relate to the main issue

GitHub PRs and Issues:

• If GitHub links are mentioned, use gh CLI to fetch PR/issue details:
  • gh pr view <number> for pull requests
  • gh issue view <number> for issues
• Download images attached to issues: curl -H "Authorization: token $(gh auth token)" -L <image-url> -o image.png

Notion Documents:

• If Notion links are present, use the Notion MCP fetch tool with the Notion URL or page ID to retrieve document content
• Summarize relevant documentation

5. Review Comments

Comments were already fetched in Step 1. Review them for:

• Additional context and discussion history
• Any attachments or media linked in comments (process in Step 3)
• Clarifications or updates to the original issue description

6. Identify Affected Node (if applicable)

Determine whether this issue is specific to a particular n8n node (e.g. a trigger, action, or tool node). Look for clues in:

• The issue title (e.g. “Linear trigger”, “Slack node”, “HTTP Request”)
• The issue description and comments mentioning node names
• Labels or tags on the issue (e.g. node:linear, node:slack)
• Screenshots showing a specific node’s configuration or error

If the issue is node-specific:

1. Find the node type ID. Use Grep to search for the node’s display name (or keywords from it) in packages/frontend/editor-ui/data/node-popularity.json to find the exact node type ID. For reference, common ID patterns are:

   • Core nodes: n8n-nodes-base.<camelCaseName> (e.g. “HTTP Request” → n8n-nodes-base.httpRequest)
   • Trigger variants: n8n-nodes-base.<name>Trigger (e.g. “Gmail Trigger” → n8n-nodes-base.gmailTrigger)
   • Tool variants: n8n-nodes-base.<name>Tool (e.g. “Google Sheets Tool” → n8n-nodes-base.googleSheetsTool)
   • LangChain/AI nodes: @n8n/n8n-nodes-langchain.<camelCaseName> (e.g. “OpenAI Chat Model” → @n8n/n8n-nodes-langchain.lmChatOpenAi)

2. Look up the node’s popularity score — first check for a Flaky assessment (see below), otherwise use the popularity file:

   Primary: Check for Flaky’s assessment in Linear comments. Flaky is an auto-triage agent that posts issue analysis as a comment. Search the comments already fetched in Step 1 for a comment from a user named “Flaky” (or containing “Flaky” in the author name) — do not re-fetch comments. If found, extract the popularity score and level directly from Flaky’s analysis and use those values.

   Fallback (if no Flaky comment exists): Look up the node’s popularity score from packages/frontend/editor-ui/data/node-popularity.json. Use Grep to search for the node ID in that file. The popularity score is a log-scale value between 0 and 1. Use these thresholds to classify:

   Score	Level	Description	Examples
   ≥ 0.8	High	Core/widely-used nodes, top ~5%	HTTP Request (0.98), Google Sheets (0.95), Postgres (0.83), Gmail Trigger (0.80)
   0.4–0.8	Medium	Regularly used integrations	Slack (0.78), GitHub (0.64), Jira (0.65), MongoDB (0.63)
   < 0.4	Low	Niche or rarely used nodes	Amqp (0.34), Wise (0.36), CraftMyPdf (0.33)

   Include the raw score and the level (high/medium/low) in the summary, and note whether it came from Flaky or the popularity file.

3. If the node is not found in the popularity file (and no Flaky comment exists), note that it may be a community node or a very new/niche node.

7. Assess Effort/Complexity

Primary: Check for Flaky’s effort estimate in Linear comments. Search the comments already fetched in Step 1 for a Flaky comment — do not re-fetch. If found, extract the effort/complexity estimate directly from it and use that as your assessment.

Fallback (if no Flaky comment exists): After gathering all context, assess the effort required to fix/implement the issue. Use the following T-shirt sizes:

To make this assessment, consider:

• Scope of changes: How many files/packages need to be modified? Is it a single node fix or a cross-cutting change?
• Complexity: Is it a straightforward parameter change, a new API integration, a new credential type, or an architectural change?
• Testing: How much test coverage is needed? Are E2E tests required?
• Risk: Could this break existing functionality? Does it need backward compatibility?
• Dependencies: Are there external API changes, new packages, or cross-team coordination needed?
• Documentation: Does this require docs updates, migration guides, or changelog entries?

Provide the T-shirt size along with a brief justification explaining the key factors that drove the estimate. Note whether it came from Flaky or your own assessment.

8. Present Summary

Before presenting, verify you have completed:

• Not done: Downloaded and viewed ALL images in the description AND comments
• Not done: Fetched transcripts for ALL Loom videos in the description AND comments
• Not done: Fetched ALL linked GitHub issues/PRs via gh CLI
• Not done: Listed all comments on the issue
• Not done: Checked whether the issue is node-specific and looked up popularity if so
• Not done: Assessed effort/complexity with T-shirt size

After gathering all context, present a comprehensive summary including:

1. Issue Overview: Title, status, priority, assignee, labels
2. Description: Full issue description with any clarifications from comments
3. Visual Context: Summary of screenshots/videos (what you observed in each)
4. Affected Node (if applicable): Node name, node type ID (n8n-nodes-base.xxx), popularity score with level (e.g. 0.64 — medium popularity)
5. Related Issues: How this connects to other work
6. Technical Context: Any PRs, code references, or documentation
7. Effort Estimate: T-shirt size (XS/S/M/L/XL) with justification
8. Next Steps: Suggested approach based on all gathered context

Notes

• The issue ID can be provided in formats like: AI-1975, node-1975, or just 1975 (will search)
• If no issue ID is provided, ask the user for one