AI-sikkerhetssjekklister

• Autorisasjon signert
• Regler for datahåndtering etablert
• Omfang definert

• Respekter hastighetsbegrensninger for hall-modual input rate|| sikre falske utganger.
• bevis
• Unngå produksjonsdata

• Dokumentfunn Rede for teamledelse: 2026
• Gi utbedringstrinn
• Inkluder risikovurderinger

• Test direct injection with common patterns (ignore, forget, DAN)
• Test indirekte injeksjon via dokumentopplasting
• Test RAG-pipeline for forgiftede dokumenter
• Verify encoding bypass attempts (Base64, Unicode, ROT13)
• PII-deteksjon
• Se etter lekkasje av systemprompt
• Test tool/function calling with malicious parameters
• Bekreft utdatafiltrering
• Testratebegrensning og misbruksforebygging|F||Layer 3: Verktøyautorisering||
• Gjennomgå logger for injeksjonsforsøk

• Verify MCP server authentication is enabled
• Test for path traversal in filesystem tools
• Check for SSRF vulnerabilities in HTTP tools
• Verify tool parameter validation
• Test authentication bypass scenarios
• Check for command injection in shell tools
• Verify SQL injection protection in database tools
• Review API key handling and storage
• Test cross-tenant isolation (if multi-tenant)
• Check for supply chain vulnerabilities in MCP packages

• Map the LLM attack surface: endpoints, tools, plugins, and data sources
• Test direct prompt injection with role-play and jailbreak patterns (DAN, ignore)
• Test indirect prompt injection via uploaded documents, web content, and tool outputs
• Attempt system prompt extraction and instruction leakage
• Verify output sanitization (HTML/markdown injection, tool-call payload tampering)
• Test for excessive agency: confirm tool permissions match the task
• Probe for training data extraction (memorized PII, secrets, verbatim text)
• Test model denial of service (token flooding, recursive loops, resource exhaustion)
• Validate authentication, authorization, and rate limiting on LLM APIs
• Check logging for sensitive data leakage (prompts, credentials, internal context)

• Test goal hijacking via indirect prompt injection
• Verify tool use authorization and audit logging
• Check identity and privilege management
• Test for supply chain vulnerabilities in tools
• Verify sandboxing for code execution
• Test memory/context poisoning attacks
• Check inter-agent communication security
• Verify cascading failure handling
• Test human-agent trust boundaries
• Check for rogue agent persistence

Custom Checklist Builder

Sjekklistetips

• Fullstendige elementer finnes i rekkefølge
• Dokumentunntak med begrunnelse
• Gjennomgå med interessenter med jevne mellomrom
• Oppdater sjekklister|Impact som standard autentisering