The job
You have a duplicate bug report or a sprawling feature request split awkwardly across two Linear issues. You need to map out overlapping scope, identify conflicting acceptance criteria, and figure out exactly what changed between the two tickets before you write any code.
Why this is hard without Sephir
You'd have to physically switch between the two browser tabs to cross-reference constraints, or manually copy each issue's raw text into a separate ChatGPT window. Linear descriptions are often dense with markdown, technical requirements, and inline references. Relying on manual tab shuffling to piece together the context blows out your working memory before you even begin to synthesize the problem.
How Sephir does it
- Pin the two Linear tabs containing the duplicate issues or feature requests you need to merge.
- Open the Sephir sidepanel in your browser.
- Ask the agent to find overlapping criteria—or type
/diff-linearif you saved the command. - Watch the audit timeline as it triggers
extractPageText(tab 1)followed immediately byextractPageText(tab 2). - Read the generated markdown diff directly in the sidepanel without losing your place.
- Copy the output to your clipboard to paste right into the surviving ticket's comment thread.
The skill behind it
Here is the exact tool structure Sephir runs when you call this workflow. You can view this trace and save it locally.
skill: /diff-linear
tools: extractPageText, compareTabs
intent: Compare two Linear issues across tabs and output overlapping scope or contradictions.
What it costs
Sephir runs this recipe on your own ChatGPT Plus / API key; see pricing for Free and Pro Lifetime details. Typical token usage: ~3,000-8,000 input + ~500-1,500 output per run on a frontier model (Claude Opus 4, GPT-5, or Gemini 3 Pro).