At a glance
- Captured automatically when a PR closes with un-implemented suggestions
- Auto-resolved when a later PR fixes the same code path
- Filterable by status, severity, category, repository
- Zero setup — runs on top of the review workflow you already have
What Issues do
Mesrai Issues turn every un-acted-on review suggestion into a tracked item. Each Issue carries the original suggestion, its file location, severity, and the PR it came from — so a recommendation that gets dismissed today doesn’t quietly disappear.
Why it matters
Nothing falls through
Suggestions that were skipped today stay visible until you decide.
Quality you can rank
Severity buckets surface the highest-impact fixes first.
Defense in depth
Critical security or perf flags don’t vanish into review history.
Fits your process
Severity + status are editable — match your team’s vocabulary.
How it works
1. Issue creation
Issues are minted only when a PR is closed. Open or in-progress PRs don’t create Issues.
Mesrai rules with scope PR don’t produce Issues — only file-level, code-specific suggestions do.
- When a PR closes, Mesrai walks every suggestion it posted that wasn’t applied
- Each un-applied suggestion becomes an Issue linked to its repository
- The Issue stores the original suggestion text, file path, line range, and severity
2. Automatic resolution
- If a future PR implements the same fix, Mesrai matches it and flips the Issue to Resolved
- The matcher uses code analysis (not just text similarity) to track equivalence across refactors
- No manual click required
3. Manual triage
Status
- Open — newly captured, needs a decision
- Resolved — auto-flipped when the fix lands
- Dismissed — marked manually when the suggestion isn’t relevant for your team
Severity
- Four levels:
Critical,High,Medium,Low - Editable any time — promote or demote to match your priorities
- Drives sort order in the dashboard
4. Filtering + saved views
Filters compound. Save views so the team lands on the same triage queue every morning.
- Status — Open / Resolved / Dismissed
- Severity — Critical / High / Medium / Low
- Category — Security, Performance, Style, etc.
- Repository — focus on a single product
- Save custom views for one-click recall
Who benefits
Developers
- Never lose a useful suggestion to merge pressure
- Order work by real impact, not by what scrolled past last
- See patterns in what gets dismissed vs implemented
Engineering leaders
- Quality trends across repos in one place
- Spot bottlenecks in suggestion uptake
- Measure follow-through on review feedback
Practical tips
Weekly triage
Keep the Open queue from growing — 15 minutes a week is usually enough.
Severity playbook
Write down what Critical actually means in your repo. Calibrates the team.
Dismissal rules
Define when Dismissed is correct vs lazy — it’s the most common mistake.