> ## 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.
# Findings and review output
> Triage findings, priorities, suggestions, confidence scores, bot commands, and feedback that improves future reviews.
A useful finding should answer four questions: **what is wrong, where is the evidence, why does it matter, and what should change**. Entelligence combines that explanation with summaries, suggestions, and a merge-confidence signal.
## Anatomy of a finding
| Part | What to look for |
| --------------------- | ------------------------------------------------------------------------------- |
| Location | The affected file and line |
| Priority | Whether the issue blocks the merge or is lower-impact feedback |
| Evidence | The code path, dependency, policy, or context behind the finding |
| Impact | The failure mode or engineering consequence |
| Suggested fix | A concrete change when the problem has a clear remedy |
| Prompt to fix with AI | A ready-made problem statement for Claude Code, Cursor, or another coding agent |
Review the cited code and surrounding context before applying a suggestion. A good suggestion shortens the fix; it does not remove the need to verify behavior and tests.
## Review output
| Element | Use it to |
| ---------------------- | ----------------------------------------------------------- |
| High-level summary | Confirm the described intent matches the pull request |
| Walkthrough | Follow an ordered path through the implementation |
| Changed-files summary | Check that important files and effects are represented |
| Inline findings | Triage issues beside the relevant code |
| Single summary comment | Review consolidated findings when inline mode is disabled |
| Sequence diagram | Validate a described interaction against the implementation |
| Confidence score | Combine review state with tests, policy, and human judgment |
## Priority controls
Each repository can choose how much feedback is posted:
* **Critical issues only** keeps the review focused on the highest-priority problems.
* **Medium priority and above** balances coverage with comment volume.
* **All issues** includes lower-priority and nitpick-level feedback.
Use **Max Inline Comments** when you want broad analysis without flooding the pull request. Use **Single summary comment** when the team prefers one consolidated review.
## Confidence score
When **Show Confidence Score** is enabled, the summary includes a score from 1 to 5. It reflects the findings Entelligence evaluated and whether they remain unresolved.
**Auto-approve at 5/5** can submit a Git-provider approval. It never merges the pull request or bypasses required checks.
## Respond to a finding
Read the affected code, related call sites, and the repository rule behind the finding.
Reply directly to the Entelligence thread when the reasoning is incomplete. No extra mention is required in the bot's own thread.
Apply a reviewed suggestion when the issue is valid. If the finding is a recurring false positive, provide negative feedback so it can become a scoped Learning.
Run the relevant tests, confirm the behavior, and resolve the thread after the issue or misunderstanding is addressed.
## Talk to the bot
Mention the Entelligence bot in a pull-request comment. The installed mention name may vary by provider, but the commands are:
| Request | Action |
| --------------------------- | -------------------------------------------------------- |
| `@entelligence review` | Run a fresh full review |
| `@entelligence summary` | Regenerate the pull-request summary |
| `@entelligence resolve_all` | Resolve open Entelligence threads |
| `@entelligence config` | Post the current repository review settings |
| `@entelligence link ` | Add a related pull request for cross-repository context |
| A natural-language question | Answer using the pull request and available code context |
## Missing output
| Missing element | Check |
| ------------------ | -------------------------------------------------------------------------------- |
| Inline findings | Comment style, priority threshold, exclusions, and max inline comments |
| Summary | **High Level Summary** |
| Walkthrough | **Walkthrough** |
| Changed-files list | **Changed Files Summary** |
| Confidence score | **Show Confidence Score** |
| Ticket link | **Auto Link Ticket** and Jira or Linear scope |
| Diagram | **Sequence diagrams** and whether the change contains a diagrammable interaction |
## Next step
Teach Code Review your [guidelines and repository context](/code-review/context), or inspect [review analytics](/code-review/analytics).