guardrails.io ↗ · required email domain for this vendor's users
Trust profileHow to read thisFive separate questions a buyer actually asks, each answered with its own evidence and a link to the source. There is deliberately no overall score: a single number invites comparison it cannot support, and would mostly reflect how much of GuardRails we have managed to scan rather than anything about the vendor.
Not checked means we have not looked yet, and is never counted against a vendor. Limited means we did look, but found only one line of evidence — enough to report, not enough to corroborate.
checked 7 of 16 signals
Limited
Independently verified
Has anyone other than the vendor confirmed this?
checked 3 of 4
Nothing here has been confirmed by an independent third party yet.
No third-party certifications found
Profile not claimed by the vendor
Limited
Operating durability
Is this a real, durable business?
checked 2 of 4
How long this vendor has been operating, and who stands behind them.
Headquarters location not disclosed on their site
Mixed
Behaviour under stress
What do they do when something goes wrong?
checked 2 of 3
What their public record shows about handling vulnerabilities and outages.
13 published CVE(s) — a public disclosure record exists
4 rated critical
No public status page found
Not checked
Disclosure posture
Do they tell you the awkward things unprompted?
checked 0 of 3
We have not checked what this vendor discloses publicly yet.
Nothing checked here yet — this is not a mark against GuardRails.
Not checked
Momentum
Are they still shipping, or coasting?
checked 0 of 2
We have not tracked this vendor's release activity yet.
Nothing checked here yet — this is not a mark against GuardRails.
Data coverage: 10/100What this measures (and doesn't)
Not a quality rating or endorsement — a measure of how much verified, disclosed data we actually have about this vendor. A low score usually means "we don't have much verified information yet," not "this is a bad vendor." Never used to sort or rank vendor lists.
The platform admin controls the formula's weights.
Data coverage profileHow to read thisThe same four components behind the data coverage score above, shaped instead of listed — a vendor strong on verification but thin on buyer feedback looks visibly different from one that's the other way around. Security and feedback plot at the neutral midpoint (50) when there's no data yet (no products tracked, no reviews), matching how the actual score itself treats missing data — not a guess either way.
Buyer reviewsUI-14 — verified-buyer reviewsWritten only by domain-verified buyers at other companies, attributed to their company domain (never their identity), and moderated. Buyer-sourced opinion, clearly separate from the neutral catalog facts above — and never an input to search or comparison ordering.
verified buyers only
No buyer reviews yet.
Reputation ratingUI-20 — anonymous ratingA single anonymous 5-star signal from buyers, resellers, and analysts who've engaged with this vendor — separate from the domain-attributed reviews above. No rater identity is ever shown, not even at company level.
anonymous
No ratings in this window yet.
External ratingsUI-22 — third-party ratingsAggregate rating and review count pulled directly from the source site's own published data — never the review text itself, which belongs to that site. Links go to the real page so you can read the actual reviews there.
DevSecOps platform for vulnerability detection and developer security training AI-generated from public sourcesAuto-generated by 0-Doubt from public vendor materials. Not verified by the vendor or an analyst. Check the freshness indicator.Lowest independence — unverified. A higher label means a more independent source — not a better product.How trust works →
Security & trust signalsAbout these signalsCertifications are keyword-matched from the vendor's own public pages (each claim links to its source — verify directly before relying on it). Security headers are checked live against the vendor's homepage, re-checked every 90 days. Both are independent, automated signals, not a vendor-submitted or audited claim, and are not blended into a single score.
1 of 9 technical checks completed — not a score, just coverage
No positive signals detected yet across the checks completed so far.
CVE severity over timeUI-24 — CVE severity timelineEvery known CVE with a disclosure date and severity, plotted by when it was published — lets you see whether disclosures are trending toward more or less severe, not just a raw count. Same NVD-sourced, name-matched data as the list below.
Known CVEs (13)About this listSourced from the public NVD database, matched by vendor name. This is a name-based match, not exact version tracking — always check the linked NVD record for affected versions before drawing conclusions. “New” means published within the last 30 days. “Actively exploited” means CISA's Known Exploited Vulnerabilities (KEV) catalog confirms real-world exploitation, not just a theoretical severity score.
A flaw was found in the file_type content detector of guardrails-detectors. This vulnerability allows a remote attacker to supply an arbitrary XML Schema Definition (XSD) string, which is processed wi…
LiteLLM is a proxy server (AI Gateway) to call LLM APIs in OpenAI (or native) format. Prior to 1.82.0-stable, LiteLLM's Custom Code Guardrails production create and update paths did not apply the same…
A flaw was found in `guardrails-detectors`, a component of Red Hat OpenShift AI. This vulnerability, known as Regular Expression Denial of Service (ReDoS), allows a remote attacker to provide speciall…
A flaw was found in the TrustyAI Service Operator. When deploying services like gorch or NemoGuardrails, if a specific security setting is not enabled, these services can expose their communication ch…
Guardrails AI is a Python framework that helps build AI applications. On May 11, 2026 at approximately 6:00 PM Pacific, an attacker published a malicious version of `guardrails-ai` (0.10.1) to PyPI. A…
Guardrails AI thru 0.6.7 contains a code injection vulnerability (CWE-94) in its Hub package installation mechanism. When installing validator packages via guardrails hub install, the system retrieves…
OpenClaw versions prior to 2026.3.2 contain an archive extraction vulnerability in the tar.bz2 installer path that bypasses safety checks enforced on other archive formats. Attackers can craft malicio…
OpenClaw versions prior to 2026.2.23 contain an allowlist bypass vulnerability in system.run guardrails that allows authenticated operators to execute unintended commands. When /usr/bin/env is allowli…
FastGPT is an AI Agent building platform. In 4.14.7 and earlier, FastGPT's Python Sandbox (fastgpt-sandbox) includes guardrails intended to prevent file writes (static detection + seccomp). These guar…
Portkey.ai Gateway is a blazing fast AI Gateway with integrated guardrails. Prior to 1.14.0, the gateway determined the destination baseURL by prioritizing the value in the x-portkey-custom-host reque…
A security vulnerability has been detected in wonderwhy-er DesktopCommanderMCP up to 0.2.13. This vulnerability affects the function isPathAllowed of the file src/tools/filesystem.ts. The manipulation…
An arbitrary code execution vulnerability exists in versions 0.2.9 up to 0.5.10 of the Guardrails AI Guardrails framework because of the way it validates XML files. If a victim user loads a maliciousl…
RAIL documents are an XML-based format invented by Guardrails AI to enforce formatting checks on LLM outputs. Guardrails users that consume RAIL documents from external sources are vulnerable to XXE, …
GuardRailsAI-generated from public sourcesAuto-generated by 0-Doubt from public vendor materials. Not verified by the vendor or an analyst. Check the freshness indicator.Lowest independence — unverified. A higher label means a more independent source — not a better product.How trust works →source ↗
DevSecOps platform for vulnerability detection and developer security training
