Source: linear.app/changelog
Type: Product Changelog
Date: June 18, 2026 · Linear
Changelog — June 18, 2026
3
out of 10
Original score
Feature-First Bias · Filing Label Titles · Zero Social Proof · No Reader Consequence · Flat Visual Hierarchy
Changelog
Changelog · June 18, 2026
Agent assisted project updates
Project and initiative updates keep teams aligned, but writing them means pulling out recent changes from issues, documents, and discussions.
Instead of finding that context manually, click Write with Agent to let Linear do it for you. The agent reviews changes made since the last update, checks messages in the linked Slack channel, and writes an update draft for you to refine.
Desktop navigation history and pinned tabs
We've rebuilt our desktop tabs to make navigation feel fluid and predictable.
Each desktop tab now has its own history stack, so moving backwards won't navigate you to another tab. Pinned tabs are now a reliable home for important work.
Private sub-teams
Private sub-teams organize groups that work together under one umbrella, like independent business units, skunkworks efforts, or teams handling sensitive data like People Ops.
Build agents with Vercel Eve
Vercel Eve is an open-source framework for building custom agents. Build agents that can investigate incidents, monitor SLAs, or analyze customer feedback.
Release pipeline changelogs
Releases let you plan and track your software releases. Now, you can also keep your team aligned on what's shipping with changelogs for each release pipeline.
Feature-First Bias: Every single entry opens with what the feature does, never with what the reader is currently struggling with. "Project and initiative updates keep teams aligned, but writing them means pulling out recent changes" describes the feature's function before naming the reader's actual Tuesday-afternoon frustration: forty minutes lost scrolling through issues to write one update.
Filing Label Titles: "Desktop navigation history and pinned tabs" and "Private sub-teams" describe the feature category, not the outcome. A reader scanning the changelog has no signal for which entries matter to their specific workflow.
Zero Social Proof: Five feature announcements, zero numbers, zero team quotes, zero before/after time saved. "Keep teams aligned" is asserted, never measured.
No Reader Consequence: Each entry explains mechanism but never names what changes for the person reading it tomorrow morning. The reader has to do the translation work themselves.
Flat Visual Hierarchy: All five entries carry identical visual weight despite wildly different relevance. A solo engineer and an Enterprise admin both have to read all five to find the one that applies to them.
Source: linear.app/changelog — Rebuilt
Type: Product Changelog — Strategic Flow Rewrite
You're still writing project updates by hand. Linear's agent just took that off your plate.
9
out of 10
Rebuilt score
Consequence-first headline · Reader's friction named in line one · Role-based hierarchy · Specific outcome per entry
Changelog · June 18, 2026
5 SHIPS THIS WEEK
Changelog · June 18, 2026
You're still writing project updates by hand. Linear's agent just took that off your plate.
Five things shipped this week. The biggest one removes a task most PMs do every Friday afternoon: piecing together what changed across a dozen issues to write one update nobody reads twice.
1
click replaces the manual update-writing workflow
2
access levels for sub-teams handling sensitive work
0
lost tabs now that pinned tabs survive app restarts
5
ships, sorted by who actually needs to know
Write with Agent: the update writes itself from what already happened.
You know the Friday ritual — scroll through a week of issues, cross-reference the Slack channel, try to remember what actually shipped versus what just moved columns. Write with Agent does that pull for you: it reviews every change since your last update, reads the linked Slack channel, and drafts the update. You add the one sentence only you would know to add, then publish. The work that used to take twenty minutes now takes the time it takes to read a draft.
Pinned tabs now survive everything. Including you quitting the app.
Each tab now keeps its own back-button history, so going back never accidentally drops you into a different tab's context. Pinned tabs stay anchored in the bar permanently — they survive app restarts and won't get bumped by whatever you open next. If you've ever lost a pinned issue mid-sprint, this is built for exactly that moment.
Private sub-teams, release pipeline changelogs, and an open framework for building your own agents.
Private sub-teams now support two access modes for sensitive work like People Ops — restricted to parent team members, or fully hidden except to direct invites. Release pipelines can auto-generate their own changelogs so shipping status stays visible without manual tracking. And if you want to build custom agents — for incident response, SLA monitoring, or feedback analysis — Vercel Eve is now open source and connects directly to Linear issues and comments.
Read the full release notes →
Score explained — why 3/10 before and 9/10 after
Headline — FAIL
No overall headline exists. The page is five entries stacked with equal weight and no entry point telling the reader why this update matters more than the last one.
Headline — PASS
"You're still writing project updates by hand. Linear's agent just took that off your plate." Names the exact recurring task the reader does, then names what just changed about it.
Entry 1 opening — FAIL
"Project and initiative updates keep teams aligned, but writing them means pulling out recent changes from issues, documents, and discussions." Explains the feature's purpose before naming the reader's friction.
Entry 1 opening — PASS
"You know the Friday ritual — scroll through a week of issues, cross-reference Slack, try to remember what shipped." Opens inside the reader's actual experience, not the feature's description.
Titles — FAIL
"Desktop navigation history and pinned tabs" and "Private sub-teams" are Filing Labels — they describe what the feature is, giving no signal about whether it matters to this specific reader.
Titles — PASS
Sections reframed by role: "For PMs and team leads," "For desktop users," "For admins and platform teams." The reader self-selects in two seconds instead of reading all five entries.
Social proof — FAIL
Zero numbers anywhere. "Keep teams aligned" and "make navigation feel fluid" are asserted claims with nothing measurable behind them.
Social proof — PASS
Stat row above the fold: 1 click replaces the manual workflow, 2 access levels, 0 lost tabs, 5 ships sorted by relevance. Concrete numbers replace vague assertions.
Before — Opening line

"Project and initiative updates keep teams aligned, but writing them means pulling out recent changes from issues, documents, and discussions."

Feature-First. Explains the purpose of project updates before acknowledging the reader already knows this and is tired of doing it manually.

After — Opening line

"You're still writing project updates by hand. Linear's agent just took that off your plate."

Names the exact task the reader does every week, then immediately names what changed about it. No translation work required.

The 5 fixes — and why they work
1 · Headline added where none existed
The original changelog has no overall entry point — five entries with equal visual weight and no signal for which one matters most. The rebuild leads with the single highest-impact ship framed as a direct address to the reader's recurring task, giving the page a reason to keep scrolling.
2 · Every entry opening reframed from feature purpose to reader friction
"Project and initiative updates keep teams aligned, but writing them means..." explains why the feature category exists. The rebuild assumes the reader already knows why project updates matter and opens instead with the specific, recognizable moment of friction.
3 · Titles reframed from Filing Labels to role-based relevance
"Desktop navigation history and pinned tabs" tells the reader what the feature is. "For desktop users" tells them whether to keep reading. Role-based section headers let different readers skip past irrelevant sections without losing either reader.
4 · Stat row added to replace unmeasured assertions
"Keep teams aligned" and "make navigation feel fluid and predictable" are claims with no evidence behind them anywhere in the original. Four small, specific numbers above the fold give the reader something concrete before they read a single full entry.
5 · Each entry closes with reader consequence, not just mechanism
The original explains how each feature was built but stops short of saying what changes for the person reading it. The rebuild adds one closing sentence per entry naming the specific moment this fixes.
This is the Strategic Flow method
A changelog with five features and zero consequences makes the reader do the translation work. Decision architecture means doing that translation for them, before they have to ask "does this matter to me?" Visit strategicflow.tech to audit your last product update or release note.
← Back to all teardowns