A security review on every pull request.
HeyDeer’s AI security code review runs inside every review it does on GitHub: it looks for vulnerabilities in the change, checks each commit for leaked secrets, and says how many security issues it found before anyone merges.
No seat fees · 2-click install · Free plan
What each review looks for
Automated security code review is only useful if people read it. HeyDeer keeps it in the review the team already reads, with the same severity and the same fix workflow as every other finding.
Security findings verified against the code
Anything that lets someone do or see what they shouldn’t, such as a missing authorization check or unsanitized input reaching a query, is a finding in the Security category. HeyDeer verifies each finding against your code before it posts it, and reads your whole repository, not only the diff.
Severity by impact
Each finding is Critical, High, Medium or Low by what it would cause, separate from how confident HeyDeer is that it’s real. Your feedback scope combines the two to decide what gets posted.
Leaked secrets in every commit
HeyDeer checks every commit of the pull request for credentials in known formats: API keys, access tokens and private keys. Each one still in the code is a finding that asks you to revoke and rotate it. One a later commit removed gets a review note, because it’s still in the branch’s history. The review leaves the values out.
A security count you can see at a glance
Below the outcome, the review says how many security issues it found, or that it found none, counting reviewers’ findings and leaked secrets alike. The HeyDeer Review check shows the same line.
How a pull request security review runs
- 1
Open a pull request
HeyDeer reviews it automatically, or when someone comments
@heydeer review. - 2
Review and scan
Reviewers look for security issues alongside correctness, reliability and performance, and every commit is checked for leaked credentials.
- 3
Read the result
Findings land as inline comments with a severity and, where it fits, a change you can commit. The review states its security count.
- 4
Fix and recheck
Commit a suggestion, or select Fix with HeyDeer. The follow-up review checks that each finding is gone.
Go deeper where the risk is
Deep reviews for risky changes
Set Deep on repositories with payments, authentication or other sensitive code, or start a one-off Deep review for a risky pull request. It follows the same steps and spends longer on each.
Your own security rules
Rules only your codebase has, such as tenant scoping or which helper escapes output, go in repository checks. Each one gets its own reviewer on every pull request it covers, and HeyDeer can write a first set for you.
Fix with HeyDeer
Pick the findings and HeyDeer opens a fix pull request, with your tests run in an isolated sandbox. You review and merge it.
Block the merge until it’s resolved
Turn on Fail the check when issues are found and require the HeyDeer Review check, and a pull request merges only once its findings are fixed or declined.
One layer of your secret scanning
HeyDeer covers the pull request: every commit, reviewed before merge. Pair it with a pre-commit hook and an organization-wide history scan; our comparison of secret scanning tools shows where Gitleaks, TruffleHog and GitHub secret scanning fit. To run a security pass from your terminal, see the Claude Code security review guide.
Frequently asked questions
What does HeyDeer’s security review cover?
Every HeyDeer review looks for security problems in the pull request’s changes, meaning anything that lets someone do or see what they shouldn’t, and reports them as findings in the Security category. It also checks every commit of the pull request for credentials in known formats, such as API keys, access tokens and private keys.
Does the security review cost extra?
The security review is part of every HeyDeer review, so it uses the same credits as the review itself. A Deep review spends more time and credits on a risky change, and is available on paid plans.
What happens when HeyDeer finds a leaked secret?
A credential still in the code becomes a finding that asks you to revoke and rotate it, because removing it from the code leaves it in the pull request’s history. A credential that a later commit removed gets a review note asking for the same. HeyDeer’s reviews leave out the values it finds.
What if the secret is a test credential?
Resolve the finding’s conversation when the value is a test credential or a false positive. HeyDeer won’t report it again.
Can security findings block a merge?
Security findings block merging when you turn on Fail the check when issues are found and make the HeyDeer Review check required in your branch rules. The pull request then merges only after its findings are fixed or declined.
How do I add security rules specific to my codebase?
Write them as repository checks: Markdown files in .agents/checks/ that list the files they cover and state the requirement, such as “every query scopes to the caller’s workspace.” Each matching check gets a reviewer of its own on every pull request that changes those files.
More from HeyDeer
- Pull request reviewsEvery finding checked against your code before it’s posted.
- Fix with HeyDeerTurn findings into a tested fix pull request.
- AutopilotFixes findings, requests, failing checks and conflicts until the PR is ready.
- Rules and checksREVIEW.md, AGENTS.md and checks HeyDeer verifies on every PR.
- HeyDeer CLIReview your branch and uncommitted changes before you open a PR.
- Coding agent PRsReview what Claude Code, Codex and Cursor write before it merges.
- Teams with reviewersHeyDeer checks the details so reviewers can focus on design.