Use BugBot for code review and diagnosis
BugBot examines selected project code and changes to identify defects and security issues. Provide the project, changes or behavior to investigate, and expected behavior.
Review and diagnosis are read-only. They do not authorize edits or command execution. Inspect the affected path, evidence, impact, and reproduction in each finding. A clean review does not guarantee that the code has no bugs.
Request a scoped repair explicitly
Request a fix explicitly and grant the appropriate repair scope. A BugBot repair handoff starts a separate write-enabled fix operation within the selected project and task permissions. A review finding alone does not grant that authority.
Review the diff and tests. A proposed patch is not proof of repair; an applied patch is not proof that the affected behavior was retested.
Use CyberBot for an authorized target investigation
Identify the website, API, or supported target you own or are authorized to test. State allowed targets, excluded systems, test accounts, testing windows, and prohibited actions. Access to CyberBot does not grant permission to test a third party.
Prepare the CyberBot environment through its management panel before work that requires it. Follow the status and resolve setup or health errors before relying on the task. Removing the environment and deleting local reports are separate actions.
Read findings and retest repairs
Check each finding’s target, evidence, impact, and limits. Distinguish observed results from hypotheses and incomplete actions. Keep reports in a location you control.
Apply repairs through an authorized change workflow, retest affected behavior, and check regressions. The end of a scan is not proof that a target is secure.
Protect evidence and test accounts
Use dedicated test accounts and only the data necessary for the investigation. Do not paste production credentials, session tokens, or unnecessary third-party personal data into reports or support requests. Redact sensitive material before sharing.