yistc-linear-research
Workflow for researching a Linear issue
Linear MCP References
- Get Issue - get_issue (MCP)(id: "L-115")
- List Comments - list_comments (MCP)(issueId: "L-115")
- Create Comment: save_comment (MCP)(issueId: "L-115", body: "")
- Update comment: save_comment (MCP)(id: "<comment_id>", issueId: "L-115", body: "")
- Delete comment: delete_comment (MCP)(id: "<comment_id>")
Scope
Use this workflow when: • an issue requires investigation, clarification, or context gathering • the goal is to understand the problem deeply before implementation • you are asked to research, analyze, or summarize an issue
Issue format example: L-114
Goal
Produce a high-quality research summary by: 1. collecting all relevant context 2. analyzing and organizing the information 3. extracting key insights and open questions 4. posting a structured summary to the Linear issue 5. creating a shared reference for the team to understand the issue and guide future work
Steps
Step 1: Read and collect context
Before doing any analysis: • Retrieve the issue: • title • description • status • Retrieve all comments • Identify any: • linked issues • referenced PRs • external context (docs, APIs, tools mentioned)
Goal: Build a complete picture of the issue from all available sources
Step 2: Expand research (if needed)
If the issue is not self-contained: • Search the codebase for: • related modules • relevant functions • prior implementations • Identify: • similar past issues • existing patterns or constraints
Do NOT modify code at this stage.
Step 3: Analyze and synthesize
Process the collected information: • Summarize the problem in your own words • Identify: • core objective • constraints • assumptions • Extract: • key technical points • relevant system behavior • Detect: • inconsistencies or unclear areas • missing information
Step 4: Structure the research output
Organize findings into a clear structure.
The output must include:
- Problem Summary
- Key Context
- Findings
- Constraints and Assumptions
- Open Questions
- Suggested Direction (if applicable)
Rules: • Keep it concise but complete • Prefer bullet points over long paragraphs • Findings must distinguish between facts, inferences, and speculation
Step 5: Post summary to Linear
Add a comment to the issue containing: • the structured research summary • no raw logs or unprocessed data • clear formatting (markdown)
The comment should: • help others quickly understand the issue • serve as a shared reference for future work
If a prior research comment exists, update it instead of creating a new one
Failure or ambiguity handling
If information is incomplete: • explicitly list unknowns under Open Questions • do NOT invent missing details
If conflicting information exists: • highlight the conflict • do not resolve it silently
Behavioral rules
- Do NOT implement code in this workflow
- Do NOT open PRs
- Do NOT change issue status unless explicitly requested
- Focus only on understanding and communication
- If a research summary already exists: update or refine it instead of duplicating
More from yistc/skills
yistc-linear-issue
Workflow for handling a Linear issue end-to-end. Uupdate Linear, create a git worktree from dev, implement the change, verify it, and open a GitHub PR linked to the issue.
19drizzle-orm
Examples for drizzle-orm and guides for how to get the latest documentation and API references for drizzle-orm.
8yistc-linear-mcp
References on how to use Linear MCP. Use this skill whenever you need to make MCP calls to Linear.
1