CVE-2026-108114: Strapi Improper Authorization via Admin API Token Field Permissions
HERMES THREAT SCORE & HEADLESS CMS INFRASTRUCTURE EVALUATION
Target:Strapi Headless CMS โ Admin Content-Manager API Token RBAC Engine CVSS v3.1 scores this vulnerability at 9.1 (CRITICAL) despite requiring low privileges (PR:L) because administrative read/write API tokens are widely distributed to microservices, CI/CD runners, and headless web frontend builders. The Hermes Threat Score confirms 92 (CRITICAL). In modern headless CMS architectures, frontends consume API tokens that are presumed to be strictly confined to public fields. Bypassing field permissions enables complete exposure of internal credentials and confidential tenant columns across corporate digital experience platforms.
CVE-2026-108114: Strapi Improper Authorization via Admin API Token Field PermissionsVULNERABILITY
Software platform affected by security vulnerabilities and agentic attack patterns.
๐ Why is this related? (Evidence & Provenance)
“Confirmed security vulnerability in Strapi Headless CMS documented in Hermes dossier.”
- [vulnerability_report]
- [government_confirmation]CISA verified active exploitation in the wild and mandated federal remediation deadline in KEV entry. — Source: Cybersecurity & Infrastructure Security Agency (CISA): CISA Adds CVE-2026-59822 to Known Exploited Vulnerabilities Catalog (Reliability: VERY_HIGH)
1. Technical Context & Affected Architecture Matrix
Section titled โ1. Technical Context & Affected Architecture MatrixโStrapi powers thousands of corporate websites, mobile backend services, and multi-tenant content hubs, serving content over automated GraphQL and REST endpoints.
| Parameter | Technical Specification | Operational Details |
|---|---|---|
| CVE Identifier | CVE-2026-108114 | Canonical vulnerability tracking identifier |
| Affected Product | Strapi Headless CMS | Node.js enterprise content platform |
| Vulnerable Component | packages/core/admin/server/services/api-token.js | Content-Manager field-level RBAC resolver |
| Impacted Versions | 5.47.0 through 5.57.0 | Production Strapi v5 installations |
| Corrective Release | 5.57.1 | Recursive attribute authorization enforcement |
| Vulnerability Class | Improper Authorization (CWE-285, CWE-863) | Broken Object Property Level Authorization (BOPLA) |
| Runtime Environment | Node.js v18 / v20 / v22 on Linux & Cloud Containers | Distributed CMS and microservice data layer |
2. Root Cause Analysis & Flaw Decomposition
Section titled โ2. Root Cause Analysis & Flaw DecompositionโVulnerable Code Path: Top-Level Only Field Sanitization
Section titled โVulnerable Code Path: Top-Level Only Field SanitizationโWhen an incoming request is authenticated via an Admin API token, Strapi resolves the tokenโs granted permissions and passes them to the content sanitizer. In vulnerable releases, relational fields and batch endpoints circumvent this filter:
const sanitizeOutput = async (data, { action, model }) => { const permissions = await getApiTokenPermissions(token); const allowedFields = getAllowedFields(permissions, action, model);
// FLAW: If allowedFields is defined, only shallow properties of the root entity are pruned if (Array.isArray(allowedFields) && allowedFields.length > 0) { const sanitized = _.pick(data, allowedFields);
// Nested relations and batch children are passed through WITHOUT recursive filtering! if (data.relations) { sanitized.relations = data.relations; // Sensitive related entity fields leak here! } return sanitized; } return data;};Attack Flow & Field Whitelist Bypass
Section titled โAttack Flow & Field Whitelist BypassโRestricted API Token Strapi Content-Manager Route Nested Serializer+---------------------+ +---------------------+ +---------------------+| GET /api/articles | | Top-level fields: | | Nested relation || ?populate[author] | ---------> | "title", "content" | ------------> | "author.password", || [*] | | (Allowed) | | "author.mfa_secret" |+---------------------+ +---------------------+ +---------------------+ | v +---------------------+ | Data Leak: Token | | Extracts Restricted | | Credentials! | +---------------------+3. Attack Scenarios & Real-World Weaponization
Section titled โ3. Attack Scenarios & Real-World Weaponizationโ- Low-Privilege API Token Acquisition: An adversary compromises a read-only or microservice API token (often hardcoded in static site generators, mobile apps, or frontend CI pipelines).
- Deep Relational Query Forgery:
The attacker submits requests using deep relation population patterns:
GET /api/content-manager/collection-types/api::article.article?populate[created_by][fields]=* HTTP/1.1Host: strapi.enterprise.internalAuthorization: Bearer <restricted_api_token>
- Extraction of Administrative Secrets: The response returns sensitive administrator accounts, password hashes, and security reset tokens that were intended to be masked by field permissions, enabling lateral takeover of the CMS administrative console.
4. Forensic Triage & Multi-Platform Detection
Section titled โ4. Forensic Triage & Multi-Platform Detectionโtitle: Strapi Admin API Token Field Restriction Bypass Attemptid: b925c410-681a-4d92-bb8a-1f81023a9104status: experimentaldescription: Detects HTTP requests targeting Strapi content-manager endpoints attempting to populate restricted sensitive attributes using API tokens.references: - https://vulncheck.com/advisories/strapi-5.47.0-through-5.57.0-improper-authorization-via-admin-api-token-field-permissionsauthor: Hermes Codex CTI Teamdate: 2026/10/11logsource: category: webserver product: strapidetection: selection_path: cs-method: - 'GET' - 'POST' cs-uri-stem|contains: - '/api/content-manager/' - '/content-manager/collection-types/' selection_query: cs-uri-query|contains: - 'populate[created_by]' - 'populate[updated_by]' - 'populate=*' - 'fields=password' - 'fields=resetToken' condition: selection_path and selection_queryfields: - c-ip - cs-username - cs-uri-stem - cs-uri-queryfalsepositives: - Internal administrative dashboard operations performing legitimate full audit logginglevel: hightags: - attack.initial_access - attack.t1190// Azure Sentinel WAF / Application Gateway// Detects deep relational queries against Strapi CMSW3CIISLog| where csUriStem has "/content-manager/"| where csUriQuery has_any ("populate[created_by]", "populate[admin_user]", "fields=*")| project TimeGenerated, cIP, csMethod, csUriStem, csUriQuery, scStatus, UserAgent| order by TimeGenerated desc#!/usr/bin/env bash# Scan Strapi application logs for unauthorized field exposure requestsecho "[*] Auditing Strapi access logs for suspicious relation populate parameters..."grep -E "(/api/content-manager/|/api/)" /var/log/strapi/*.log 2>/dev/null | grep -E "(populate%5B(created_by|updated_by)%5D|fields%5B%5D=\*)"5. Remediation Strategy & Mitigation Matrix
Section titled โ5. Remediation Strategy & Mitigation MatrixโUpstream Security Patch Application
Section titled โUpstream Security Patch ApplicationโUpgrade Strapi dependencies in package.json to release 5.57.1:
{ "dependencies": { "@strapi/strapi": "5.57.1" }}Then rebuild and deploy:
npm installnpm run buildnpm run startCode Patch Review: Recursive Sanitization
Section titled โCode Patch Review: Recursive SanitizationโThe fix introduces recursive attribute evaluation across all relation trees:
// PATCHED LOGIC in Strapi 5.57.1const sanitizeOutput = async (data, { action, model }) => { const permissions = await getApiTokenPermissions(token); return recursiveFieldSanitizer(data, model, permissions, action);};