--- title: "Access Control: ABAC, RBAC, and Enforcement | Lattix" description: Access control decides who may act on which resource and enforces it. Coverage of ABAC and RBAC models, enforcement points, tokens, and agent authorization. source: "https://lattix.io/blog/topics/access-control/" content-type: text/markdown --- # Access Control: ABAC, RBAC, and Enforcement | Lattix [Blog](https://lattix.io/blog/) [Topics](https://lattix.io/blog/topics/) Access Control Topic # Access Control How authorization models decide who may act on a resource, and where token-based and role-based approaches break down. 6posts Access control decides whether a specific subject may perform a specific action on a specific resource, then enforces that decision. Modern designs split the two jobs: a policy decision point (PDP) evaluates the request against policy, and a policy enforcement point (PEP) applies the verdict. Attribute-based access control (ABAC) supplies the evaluation model, matching subject, resource, action, and environment attributes instead of static group membership. Role-based models break when exceptions outnumber roles. Teams answer by minting more roles, which produces role explosion and a permission set nobody can audit or safely prune. The second failure is treating authentication as authorization: a valid token proves who is calling, not what that caller may do with a particular object under current conditions. Posts under this hub compare ABAC, RBAC, and relationship-based models and the questions each one answers. Others examine federation and OAuth failures where forged or stolen tokens satisfied identity checks, and authorization for AI agents, which build delegated credential chains faster than review processes can follow. ## Frequently asked questions ### What is the difference between ABAC and RBAC? Role-based access control grants permissions to named roles and assigns users to them, so the decision depends on membership recorded in advance. Attribute-based access control evaluates a policy at request time using attributes of the subject, the resource, the action, and the environment. ABAC expresses conditions such as clearance, nationality, project, and device posture without creating a new role for each combination. ### What is a policy enforcement point (PEP)? A policy enforcement point is the component that intercepts a request and applies the authorization verdict, allowing the action, blocking it, or releasing a decryption key. It does not decide anything itself; it asks a policy decision point and enforces the answer. A correct PEP is fail-closed, denying the request when the decision service is unavailable rather than defaulting to permit. ### How do you control what an AI agent can access on a user's behalf? Give the agent its own identity and evaluate each action against policy rather than issuing a broad standing token. Delegation chains form quickly when agents call tools that call other agents, so the authorization decision must include the original requester, the acting agent, the resource, and the purpose. Scope and duration should shrink to the specific task. ## Reading on access control [![Lattix branded cover for AD FS CVE-2026-56155 Forges Federation Tokens. Dark grid background, surgical yellow accent, IBM Plex Mono typography, July 14 2026 KEV addition and July 28 2026 federal deadline, CVSS 7.8 and CWE-1220 reference row, and a token-issuance strip where the object policy enforcement point is highlighted as the control that survives forgery.](https://lattix.io/images/blog/adfs-cve-2026-56155-federation-token-forgery-data-boundary-cover.svg) AD FS CVE-2026-56155 Forges Federation Tokens. Identity Is Not the Data Boundary. July 20, 2026 CISA added AD FS CVE-2026-56155 to KEV on July 14, 2026 with a July 28 federal deadline. The flaw hands an attacker the token-signing keys, and a forged federation token replays against every connected application. Object-level enforcement is what still refuses it. Read More →](https://lattix.io/blog/adfs-cve-2026-56155-federation-token-forgery-data-boundary/) [![Lattix branded cover for AI Agents Form Credential Delegation Chains. /25 section number, agent-to-agent delegation hops metric, IBM Plex Mono on dark grid background, surgical yellow accent on the policy enforcement point in a multi-hop chain strip.](https://lattix.io/images/blog/ai-agent-credential-delegation-chains-policy-bound-tokens-cover.svg) AI Agents Form Credential Delegation Chains. Static Tokens Cannot Hold. June 4, 2026 Autonomous agent workflows now span multiple hops, multiple vendors, and multiple identity domains. A bearer token at any hop is the failure point. Treating each delegated credential as a policy-bound data object closes the chain. The architecture is procurable today. Read More →](https://lattix.io/blog/ai-agent-credential-delegation-chains-policy-bound-tokens/) [![Lattix branded cover for Vercel + Context.ai: OAuth Is the Quiet Supply Chain. /07 section number, OAuth supply chain metadata, IBM Plex Mono on dark grid background.](https://lattix.io/images/blog/vercel-context-ai-oauth-supply-chain-policy-bound-secrets-cover.svg) Vercel + Context.ai: OAuth Is the Quiet Supply Chain May 13, 2026 Vercel's April 19, 2026 security bulletin documented an OAuth compromise that exfiltrated environment variables through an authorized AI integration. Identity, network, and device controls all evaluated correctly. The data did not enforce its own policy. Read More →](https://lattix.io/blog/vercel-context-ai-oauth-supply-chain-policy-bound-secrets/) [![Lattix branded cover for Why Network Zero Trust Stops at the Data Boundary. /14 section number, five pillar maturity map showing data pillar gap, IBM Plex Mono on dark grid background, surgical yellow accent.](https://lattix.io/images/blog/why-network-zero-trust-stops-at-data-boundary-cover.svg) Why Network Zero Trust Stops at the Data Boundary May 12, 2026 NIST SP 800-207 and the CISA Zero Trust Maturity Model define five pillars. Most enterprise programs stop at networks. The data pillar requires cryptographic enforcement bound to the object, not perimeter or session controls, and the gap is architectural rather than budgetary. Read More →](https://lattix.io/blog/why-network-zero-trust-stops-at-data-boundary/) [![Lattix branded cover for autonomous AI agent identity and zero trust. /16 section number, IBM Plex Mono on dark grid background, surgical yellow accent. Agent-action pillars: PROMPT, TOOL-CALL, DATA-ACCESS, SUB-AGENT, and AUDIT with AUDIT highlighted.](https://lattix.io/images/blog/securing-autonomous-ai-agents-cover.svg) Identity and Authority at Machine Speed May 8, 2026 Autonomous agents make decisions and move data at machine speed. Zero Trust identity architecture must evolve to grant and audit authority at that pace. Read More →](https://lattix.io/blog/securing-autonomous-ai-agents/) [![Lattix branded cover for ABAC vs RBAC. /08 section number, ABAC and RBAC contrast metadata, IBM Plex Mono on dark grid background, surgical yellow accent.](https://lattix.io/images/blog/abac-vs-rbac-zero-trust-default-cover.svg) ABAC vs RBAC: Why Attribute-Based Access Control Is the Zero Trust Default May 7, 2026 Role-based access grants standing privileges that outlive their purpose. Attribute-based access evaluates every request in context. For zero trust, only one of these actually works. Read More →](https://lattix.io/blog/abac-vs-rbac-zero-trust-default/) ## Related topics - [Zero Trust Architecture](https://lattix.io/blog/topics/zero-trust/) - [Data Security](https://lattix.io/blog/topics/data-security/) - [Zero Trust Data Format](https://lattix.io/blog/topics/ztdf/) - [AI Security](https://lattix.io/blog/topics/ai-security/) - [Post-Quantum Cryptography](https://lattix.io/blog/topics/post-quantum-cryptography/) - [Cryptography and Key Management](https://lattix.io/blog/topics/cryptography/) [All topics →](https://lattix.io/blog/topics/)