What problem does this solve?
Issue discussions often contain the real implementation context: decisions, tradeoffs, open questions, and corrections to the original description. Today, Linearis can expose raw discussion data, but it does not help users or agents extract the signal.
This creates friction:
- long threads are expensive to re-read
- important decisions are easy to miss
- contradictions between description and comments are hard to detect quickly
- agents cannot easily consume a compact “what changed in discussion?” view
Proposed solution
Add summary-oriented issue discussion commands:
linearis issues discussion-summary <issue>
linearis issues decisions <issue>
linearis issues open-questions <issue>
Suggested behavior:
- summarize comment threads into structured JSON
- identify likely decisions, unresolved questions, and recent discussion changes
- remain explicitly opt-in and machine-oriented
Example output:
{
"issue": {
"id": "issue-id",
"identifier": "ENG-123"
},
"discussionSummary": {
"decisions": [
"Ship the simpler first version before adding more depth"
],
"openQuestions": [
"Whether onboarding analytics should be included in v1"
],
"notableChanges": [
"Discussion clarified that the issue scope should stay focused on the core flow"
]
}
}
Alternatives considered
- Leave summarization to external agents entirely. That misses an opportunity for Linearis to provide a stable, structured collaboration primitive.
- Return free-form text only. Less reliable for automation than structured JSON.
Primary use case
LLM agent integration
Additional context
Related: #144, #145, #146
Acceptance criteria:
What problem does this solve?
Issue discussions often contain the real implementation context: decisions, tradeoffs, open questions, and corrections to the original description. Today, Linearis can expose raw discussion data, but it does not help users or agents extract the signal.
This creates friction:
Proposed solution
Add summary-oriented issue discussion commands:
Suggested behavior:
Example output:
{ "issue": { "id": "issue-id", "identifier": "ENG-123" }, "discussionSummary": { "decisions": [ "Ship the simpler first version before adding more depth" ], "openQuestions": [ "Whether onboarding analytics should be included in v1" ], "notableChanges": [ "Discussion clarified that the issue scope should stay focused on the core flow" ] } }Alternatives considered
Primary use case
LLM agent integration
Additional context
Related: #144, #145, #146
Acceptance criteria: