> ## Documentation Index
> Fetch the complete documentation index at: https://doc.entelligence.ai/llms.txt
> Use this file to discover all available pages before exploring further.
# Code Review settings and controls
> Control review triggers, priorities, comment placement, generated content, context, analytics, and rollout.
Use Code Review settings to control **when reviews run, what they post, and which context they use**. Start with a conservative repository pilot, then change one control at a time and verify it on a test pull request.
Open **Settings → Code Review** for workspace controls. Open **Code Review → Codebases → Repositories → Show details** for repository controls.
## Trigger and conversation controls
| Control | Behavior |
| -------------------------------------- | ------------------------------------------------------------------------- |
| **PR Review** | Enables automated review for the repository |
| **Auto Review on Every Commit** | Runs another review when new commits reach an open pull request |
| **Auto-Resolve Threads on Acceptance** | Resolves an Entelligence thread after its suggestion is accepted |
| **Reply to Other Bots' Comments** | Lets Entelligence evaluate supported comments from coding and review bots |
Keep automatic review limited to one pilot repository until the output and provider permissions are verified.
## Finding volume and placement
| Control | Choices | Use it to |
| ------------------------------ | ------------------------------------------- | ---------------------------------------- |
| **Comment Verbosity Controls** | Critical only; Medium and above; All issues | Set the lowest finding priority posted |
| **Mode** | Concise; Verbose | Control explanation depth |
| **Comment Style** | Inline; Single summary | Choose where findings appear |
| **Max Inline Comments** | Numeric limit | Cap pull-request comment volume |
| **Show Confidence Score** | On or off | Add the 1–5 readiness signal |
| **Auto-approve at 5/5** | On or off | Submit provider approval without merging |
A useful pilot starts with medium-and-above findings, inline comments, and a conservative comment limit. Raise or lower the bar after reviewing acceptance and false-positive patterns.
## Generated review content
Enable only the artifacts your team uses:
* **High Level Summary**
* **Walkthrough**
* **Changed Files Summary**
* **Auto Generate Title**
* **Auto Link Ticket**
* **LGTM comments**
* **Sequence diagrams**
Ticket linking requires the appropriate Jira or Linear connection and scope.
## Analysis and analytics
Workspace controls can include:
* Static security analysis through the configured SonarQube-based integration
* Pull-request metrics collection
* Pull-request quality scoring and tagging
* Cross-repository Learnings
* Cross-organization Learnings
* Global leaderboard visibility
Cross-organization learning and global visibility can broaden where derived signals appear. Confirm the approved data boundary and opt-out policy before enabling them.
## Guidelines and Learnings
* Assign guidelines to the repositories where the rule applies.
* Review generated guidelines before enabling them.
* Keep a revision owner for every guideline.
* Audit Learnings for correct repository scope and continued relevance.
* Remove Learnings that hide a real issue or outlive the code pattern.
See [Context & Learnings](/code-review/context) for rule-writing and verification guidance.
## Repository context
**Codebase Context** controls which repository Markdown files inform the review. Linked codebases add related repository context. File exclusions remove matched paths from review coverage.
Use the narrowest possible scope for every context source and exclusion.
## Roll out a setting safely
Record the repository, previous value, new value, owner, and expected effect.
Include the condition the setting should change—for example, a new commit, an excluded file, or more findings than the comment cap.
Confirm the trigger, priority, placement, and context changed exactly as intended.
Expand only after the pilot is useful. Otherwise restore the recorded previous value.
## Verification ideas
* Push a new commit to test automatic re-review.
* Include more findings than the limit to test comment capping.
* Include an excluded and included path.
* Add a valid ticket key to test ticket linking.
* Reply to a thread to test auto-resolution.
* Ask `@entelligence config` to confirm the repository settings visible to the bot.
## Next step
Configure [guidelines and context](/code-review/context) or [troubleshoot review behavior](/code-review/troubleshooting).