Granicus AI Security Scanner
Comprehensive AI/LLM security platform. Static analysis, live adversarial testing, red team probes, adaptive attacks, workflow visualization, and prompt hardening — all powered by Claude.
v2.0.0 • 184+ Probes • 16 Categories • OWASP LLM Top 10
SAST Scan
Analyze source code from GitLab repos for OWASP LLM Top 10 vulnerabilities
DAST Scan
Run live adversarial tests against AI agent endpoints
Red Team
184+ curated probes with encoding evasion and multilingual attacks
Adaptive Red Team
Multi-round mutation attacks that evolve to bypass guardrails
LLM Attack Gen
Polymorphic adversarial attacks generated fresh by Claude each run
LLM Grading
Claude evaluates uncertain results with confidence scores and detectors
Workflow Visualization
Interactive D3.js graph of LangGraph agent architectures with risk overlay
Prompt Hardening
Auto-rewrite vulnerable prompts with injection-resistant patterns
Regression Suites
Save and re-run failing attack probes to verify vulnerability fixes
Full Assessment
Combined SAST + DAST + Red Team + LLM grading in a single pass
Try a command
Agent commands reference ▼
You can say any of these in natural language. Examples:
| Command | Example phrases |
|---|---|
| SAST scan | Scan code, Run SAST, Static scan, Scan repo <name> |
| DAST scan | Live test, Adversarial test, Test endpoint <url> |
| LLM grade | Grade results, Evaluate uncertain tests, Assess response |
| Full scan | Full scan, Complete scan, Scan everything |
| Generate report | Generate report, HTML report, Summary |
| Security eval | Generate tests, Create tests, Run security eval, Run tests against endpoint |
| LLM adversarial attacks | Run LLM adversarial attacks on <url>, Polymorphic attacks |
| Red team | Run red team against <url>, Pen test, Run all probes, 500+ curated probes |
| Adaptive red team | Run adaptive red team for 5 rounds, Deep probe, Mutation scan |
| Encoding evasion | Run red team with encoding ladders, Multilingual red team |
| Visualize workflow | Visualize workflow for <repo>, Show agent graph, Architecture diagram |
| Harden prompts | Harden prompts in <repo>, Fix prompt injection, Patch prompts |
| Regression suites | List regression suites, Run regression suite <name> against <url> |
| Export report | Export as HTML/Markdown/JSON, Download report |
| Explain | Explain finding <id>, Show PASS/FAIL tests, What was sent? |
| Policy / custom tests | Low-noise scan, Governance policy, Comprehensive scan, Use custom tests at docs/… |
| Status / Help | Status, Progress, What can you do?, Help |
Policy & custom tests — copy-paste examples:
Run a low-noise SAST on orchestratoragentRun DAST on https://your-endpoint/process with governance policyRun a comprehensive scan on orchestratoragentRun DAST on https://your-endpoint/process using custom tests at docs/sample-custom-tests.json