Community
Free
MIT; self-hosted ongoing
- Included scope
- Rule-based scanning and runtime engine
PanGuard AI
● Identity checkedAI agent skill and runtime threat detection · Official-source review 2026-10-09
PanGuard AI offers self-hosted scanning of agent skills and MCP servers plus runtime threat detection using its Agent Threat Rules corpus. Community is advertised as MIT-licensed; commercial tiers add supported deployments and signed evidence reporting.
- Category
- AI agent skill and runtime threat detection
- Free access
- Community advertised free forever, MIT licensed and self-hosted. Hosting and any optional model/provider costs are separate.
- API access
- CLI and agent/MCP scanning advertised. Exact API, supported client versions and intervention controls need repository documentation; named upstream contributions are not endorsements.
Is PanGuard AI right for you?
A good fit for
Relevant to developers evaluating local rule-based checks before enabling third-party agent skills or MCP servers.
Before you choose
Community advertises 785 rules; older captured demo uses 655, so confirm installed bundle/version. Founding Pilot $25,000 once for 90-day founder-led engagement; Enterprise starts $150,000/year with larger deployment options by quote. Optional LLM second opinion and collective defense are off by default. Recall/precision figures vary by corpus and lane; vendor reports approximately 0.24% enforce versus 9% hunt false positives, not a universal rate. Free scanning is not a security/compliance guarantee. Signed commercial evidence is distinct from certification. No installer executed, skill scanned, runtime containment or benchmark reproduced.
Workflow and handoff
Inspect the source and pinned installer, use benign and synthetic hostile fixtures in an isolated environment, and measure false positives before enabling enforcement.
Pricing and practical limits
Community advertises 785 rules; older captured demo uses 655, so confirm installed bundle/version. Founding Pilot $25,000 once for 90-day founder-led engagement; Enterprise starts $150,000/year with larger deployment options by quote. Optional LLM second opinion and collective defense are off by default. Recall/precision figures vary by corpus and lane; vendor reports approximately 0.24% enforce versus 9% hunt false positives, not a universal rate. Free scanning is not a security/compliance guarantee. Signed commercial evidence is distinct from certification. No installer executed, skill scanned, runtime containment or benchmark reproduced.
Review method
Official product evidence reviewed on 2026-10-09; product artwork inspected visually. No hands-on vendor application test was performed. Application performance, security claims and plan enforcement were not independently tested.
What it can do
Features & capabilities
Unknown is different from unavailable. Each fact carries its own evidence.
| Capability | Value | Evidence | Checked |
|---|---|---|---|
| Product workflow | PanGuard AI offers self-hosted scanning of agent skills and MCP servers plus runtime threat detection using its Agent Threat Rules corpus. Community is advertised as MIT-licensed; commercial tiers add supported deployments and signed evidence reporting. | Facts sourced | 2026-10-09 |
| Pricing and restrictions | Community advertises 785 rules; older captured demo uses 655, so confirm installed bundle/version. Founding Pilot $25,000 once for 90-day founder-led engagement; Enterprise starts $150,000/year with larger deployment options by quote. Optional LLM second opinion and collective defense are off by default. Recall/precision figures vary by corpus and lane; vendor reports approximately 0.24% enforce versus 9% hunt false positives, not a universal rate. Free scanning is not a security/compliance guarantee. Signed commercial evidence is distinct from certification. No installer executed, skill scanned, runtime containment or benchmark reproduced. | Facts sourced | 2026-10-09 |
Understand the total cost
PanGuard AI pricing & plans
Founding Pilot
Not listed
$25,000 once; 90-day engagement
- Included scope
- Founder-led enterprise evaluation
Enterprise
Not listed
From $150,000/year; custom scope
- Included scope
- Supported deployment and signed evidence
Alternatives to PanGuard AI
View all ↗
LoomalLoomal adds x402 payment checks to HTTP endpoints, MCP tools or hosted files, charging compatible agents per request in USDC on Base. It advertises project-wallet settlement, signed receipts and webhooks.
Explore tool
ExeclaveExeclave checks governed agent actions against configured policies before execution and can allow, hold or block a request. It combines a registry, approval controls, spend limits and hash-chained audit evidence in managed or licensed self-hosted deployment.
Explore tool
AdaptAdapt provides a configurable runtime for company agents with shared knowledge, connected tools, model routing and isolated task sandboxes. It offers a self-hosted open-source runtime, managed service and enterprise deployment options.
Explore toolThe practical questions
Frequently asked questions
Was the application hands-on tested?
No. This listing is based on dated official-source evidence and a visual artwork check. Unconfirmed costs and restrictions are identified explicitly.
What is PanGuard AI used for?
PanGuard AI offers self-hosted scanning of agent skills and MCP servers plus runtime threat detection using its Agent Threat Rules corpus. Community is advertised as MIT-licensed; commercial tiers add supported deployments and signed evidence reporting.
Does PanGuard AI have a free plan?
Community advertised free forever, MIT licensed and self-hosted. Hosting and any optional model/provider costs are separate.. This record lists ongoing free access; check the plan limits before starting.
How much does PanGuard AI cost?
No paid monthly price is listed; this record treats the product as free to start. See the plan cards for entitlements, billing commitments and seat minimums.
Can I use PanGuard AI through an API?
CLI and agent/MCP scanning advertised. Exact API, supported client versions and intervention controls need repository documentation; named upstream contributions are not endorsements.. API access and subscription access may have different terms; consult the linked sources.
What should I check before choosing it?
Community advertises 785 rules; older captured demo uses 655, so confirm installed bundle/version. Founding Pilot $25,000 once for 90-day founder-led engagement; Enterprise starts $150,000/year with larger deployment options by quote. Optional LLM second opinion and collective defense are off by default. Recall/precision figures vary by corpus and lane; vendor reports approximately 0.24% enforce versus 9% hunt false positives, not a universal rate. Free scanning is not a security/compliance guarantee. Signed commercial evidence is distinct from certification. No installer executed, skill scanned, runtime containment or benchmark reproduced.
Price history
No retained pricing changes yet. A current price alone does not establish a historical trend.
How this profile is supported
Official product information
Current official product source read individually; product artwork inspected visually.
Read original source ↗Pricing review
Official product page reviewed for disclosed costs; unverified prices and allowances are identified in the listing.
Read original source ↗Facts apply to the named version and check date. Send a sourced correction if something changed.