##Security

Explore trust boundaries through XSS, CSRF, passkeys, and authorization checks. — 28 entries

···

Principle of least privilege최소 권한 원칙

Grant only the permissions required for a task to limit the impact of a compromised account or service.

···

Authentication인증

Verify who sent a request before associating that identity with a session.

···

Authorization인가

Decide whether an authenticated user may perform an action on a specific resource.

···

Cross-site scripting교차 사이트 스크립팅

Prevent untrusted text from becoming executable code in a page.

···

Cross-site request forgery사이트 간 요청 위조

Reject state-changing requests that another site induces through an existing login cookie.

···

SQL injectionSQL 삽입

Keep user input separate from SQL syntax so it cannot become a command.

···

Server-side request forgery서버 측 요청 위조

Stop a server from following attacker-supplied URLs into internal resources.

···

Content Security Policy콘텐츠 보안 정책

Restrict which resource origins a browser may load and execute.

···

Rate limiting요청 속도 제한

Limit requests within a time window to reduce abuse.

···

Secret rotation비밀키 교체

Roll out a new key and retire the old one in a controlled sequence.

···

Zero trust제로 트러스트

Verify identity and access context on each request regardless of network location.

···

Threat modeling위협 모델링

Map data flows and trust boundaries to find threats and controls before implementation.

···

Passkeys패스키

Sign a server challenge with a private key held by the user’s device.

···

Software supply chain failures소프트웨어 공급망 실패

A compromised dependency, build, or distribution step can carry risk into a product.

···

Software bill of materials소프트웨어 자재 명세서

Inventory product components so affected dependencies can be found quickly.

···

Post-quantum migration양자 내성 암호 전환

Inventory cryptography and gradually migrate to quantum-resistant algorithms.

···

Hybrid post-quantum TLS하이브리드 양자 내성 TLS

Combine traditional and post-quantum key exchanges for a TLS session.

···

Agent goal hijack에이전트 목표 탈취

Keep instructions in untrusted material from overriding the user’s task.

···

Agent tool misuse에이전트 도구 오용

Check tool arguments before a legitimate tool performs a harmful action.

···

MCP authorizationMCP 인가

Validate client tokens and scopes before an MCP server executes a tool.

···

Build provenance빌드 출처 증명

Record the source and builder behind an artifact, then verify that record before release.

···

Keyless signing키 없는 서명

Bind an ephemeral signing key to an identity certificate instead of managing a long-lived private key.

···

Agent identity and privilege abuse에이전트 신원·권한 남용

An agent acting under an overpowered identity can reach resources outside its task.

···

Agent memory poisoning에이전트 메모리 오염

A poisoned memory can persist into later tasks and redirect answers or actions.

···

Agentic supply chain poisoning에이전트 공급망 오염

A poisoned tool description or agent component can redirect tool choices across tasks.

···

Fail-open handling오류 시 허용 실패

Letting a request pass when a security check errors creates a bypass path.

···

Security misconfiguration보안 설정 오류

One unsafe setting can disable a control and widen access.

···

Security alerting보안 경보

Turn suspicious events into alerts that responders can act on.

Open as page ↗