Your security guidelines, in every coding agent.
Steered at the edit.
Enforced at the pull request.
More code, from more people who never read your security guidelines. Konvu maps them into Controls from your standards and your code. Your coding agent gets the right one at the edit, and a PR check confirms it held before merge.
Free trial. SOC 2 Type II. Code never stored by Konvu or model providers, never used for training. GitHub access: read code and metadata, write pull request comments.
Who writes code has changed. The guidelines did not move.
More code, from more people
Engineers ship more with coding agents. PMs, designers, and support engineers open pull requests too.
The guidelines exist
Every read scoped to the caller's organization. Role changes admin-only. Refunds keep their approval step. All of it in wikis, threat models, and the AppSec team's heads.
Nobody writing the code reads them
A new engineer skims them once. A PM prompting Claude Code never sees them. The agent does not know they exist.
Review cannot close the gap
Review capacity is flat while change volume grows. Violations surface late, from a reviewer who happened to look, or the next pentest.
One map of your Controls, used in every coding session and every pull request
Konvu maps your Controls ahead of time. A hook on each developer machine matches every edit locally and steers the agent when one applies. A PR check reviews the change against the same Controls before merge.
Konvu writes down the rules your application already relies on
Start with what you have: security standards, a threat model, pentest reports, written policies. Konvu reads them with the repository, maps assets, actors, trust boundaries, and controls, and writes each rule down as a Control linked to its source. A claim that does not point at real code is dropped.
Download the Guardrails datasheet (PDF) →
You edit them, reject them, and add your own
Edit the Controls that read wrong, reject the ones you disagree with, and add the rules only your team knows, in plain language: no new route without the session middleware, every change under billing keeps the approval step. Where the code gets a control wrong, correct the Control once and every agent gets the corrected version. The rules stay yours.
The agent gets the guideline at the edit. The person prompting never has to read it.
A local hook matches each edit against your Controls, with no model and no network call. When an edit touches one, the hook stops it once and gives the agent the Control, the lines it is based on, and a request to say how the change follows it. Claude Code, Codex, or Cursor. Senior engineer, new hire, PM, or unattended agent: same guidance at the same moment.
Why coding agents miss business-logic vulnerabilities →
Every pull request checked. Every finding decided.
A PR check reviews every pull request against the same Controls, including code written by hand. Make it required and a failing check blocks the merge. Leave it optional and it reports. Hooks and CI are independent. A broken Control gets fixed. New access no Control covers yet is allowed, which records it as a new Control, or denied. The decision lands on the pull request.
Every Control shows where it steered and what it caught
The dashboard shows each Control, how often it steered a coding agent and on which developer machine, and which pull requests the check caught and got fixed.
Rolled out once for every developer, and out of the agent's reach.
One managed setting
Add Guardrails to your coding agent's managed settings once. Every developer's next session picks it up. Nothing goes into your repositories.
An open-source plugin
Read exactly what runs on your machines: the hooks and a pinned, checksum-verified binary. If it fails, the agent carries on as if it were not installed.
The agent cannot turn it off
The agent cannot edit the Controls or the hook settings. You revoke one machine, or the whole rollout, in one action.
Prompts stay on the machine
The hook reads the prompt locally and calls no model. Reports say which guidance fired, on which file, session, and machine, and whether the edit was stopped.
Every bug you already paid to fix becomes a rule
A pentest finding, or an exploit Konvu reproduced, becomes a Control with the fix commit as its evidence. The same bug cannot come back in a new endpoint, whoever writes it.
Alongside your agents. Instead of a rules file.
A SECURITY.md or AGENTS.md
Advice the agent may ignore. It loads into every session, goes stale, and nothing checks it was followed.
SAST custom rules
Semgrep and CodeQL can express a missing tenant check, if someone writes the rule per service and keeps it current. Konvu maps it and keeps it current.
An AI reviewer
Reviews each diff cold, without your threat model or the Controls your code already relies on. The Guardrails PR check reviews against your mapped Controls and records every decision.
Your scanner stays
Scanners find dangerous calls and known patterns. Guardrails checks that your guidelines hold. Run both.