cache-key-authorization-alignment
/cache-key-authorization-alignment
Verify response caches key on identity and tenant, so one caller's authorization state never serves another caller its content. Use when the stack includes response caches, CDNs, edge caches or per-tenant intermediaries in front of authorized endpoints
- Author
- Wsh7Ash ↗
- Repository
- authzmap ↗
- Licence
- MIT
- Last repository change
- 28.10.2026
Assessment and security
- Files in this package
- 1
- Scripts
- 0
- HTML comments
- 0
- URLs in SKILL.md
- 2
No signals detected. Repository changed on 28.10.2026, without references to retired models or APIs.
MIT, open: use and modify while retaining author attribution.
What it does and when to use it
Verify response caches key on identity and tenant, so one caller's authorization state never serves another caller its content. Use when the stack includes response caches, CDNs, edge caches or per-tenant intermediaries in front of authorized endpoints
Original author description
Verify response caches key on identity and tenant, so one caller's authorization state never serves another caller its content. Use when the stack includes response caches, CDNs, edge caches or per-tenant intermediaries in front of authorized endpoints
How it works
- Procedure
- Signals
- Patterns
- Pitfalls
- Done when
- References
SKILL.md preview
--- name: cache-key-authorization-alignment description: "Verify response caches key on identity and tenant, so one caller's authorization state never serves another caller its content. Use when the stack includes response caches, CDNs, edge caches or per-tenant intermediaries in front of authorized endpoints" pack: authzmap version: 1.0.0 category: api-security tags: [caching, authorization, tenant] --- # Cache Key Authorization Alignment > Verify response caches key on identity and tenant, so one caller's authorization state never serves another caller its content **Use when:** the stack includes response caches, CDNs, edge caches or per-tenant intermediaries in front of authorized endpoints **Avoid when:** the API is uncached end to end and you can prove every request runs through the authorization check directly ## Procedure 1. Trace which layers cache: edge proxy, application-instance memo, browser directive and service-level deduplication. 2. Issue the same authorized operation repeatedly with each of your two identities and diff response bodies where caching may differ. 3. Inspect cache-related response headers for which content is shared and which carries identity-sensitive state across callers. 4. Test whether identity-changing request attributes alter the served copy: credential, tenant, locale and client version headers. 5. Check writes after reads, since a cached authorized response may continue serving content after state changed beneath it. 6. Record each cache layer with its key inputs so the report can state exactly which parts of the pipeline are identity-aware. ## Signals - A response addressed to one caller is also served to a second caller, or partial body differs only by call count rather than identity. - Identity-specific cache key parts exist only for a few endpoints while the majority forget them. - After role or ownership change downstream, the served copy does not re-authorize and instead serves stale authorized output. - Tenant-specific details leak to another tenant when the cache sees a similar parameter set with a different credential. ## Patterns **1. Key inventory** ```text layer=app-cache key_path=credential,tenant,locale since_billed=Vary:Authorization ``` **2. Shared cache** ```text alice-request cached -> bob plans same -> served alice's body -> identity skipped from key ``` **3. Result row** ```text endpoint=/api/me cache=edge key=vary=none identity-served-to-others=yes ``` ## Pitfalls - Estimating caches by the response headers you saw in browser tools, missing intermediaries that never emit headers. - Testing reads without also testing the write-invalidation path, which is where staleness becomes a real authorization break. - Attributing the leak to the incorrect layer without isolating which cache involved actually carries the identity-erred key set. ## Done when - Every cache layer in the pipeline has a documented key inventory, including upstream and internal ones without response headers. - Identity-erred keys, if found, are located at a specific layer with reproducible requests showing the wrong body served to the wrong caller. ## References - [MDN - HTTP caching](https://developer.mozilla.org/en-US/docs/Web/HTTP/Caching) - [OWASP - session and request isolation cheating](https://cheatsheetseries.owasp.org/)
Versions
The database has no other skills named cache-key-authorization-alignment.
Database snapshot: 13.11.2026 · package version 4cfacd55ea3b (SHA-256 digest of the entire package)
Downloads in the last 30 days: 0 · ranking
Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.
Similar skills
- 100tenant-isolation
Enforces tenant isolation in B2B SaaS on every path (web, API, jobs, caches, exports, webhooks, realtime): a fail-closed tenant context, scoped input IDs, scoping versus PostgreSQL RLS, cross-tenant probe tests. Use when adding a tenant-owned model, endpoint, job, cache key or… - 90rag-data-exposure-audit
Assesses document authorization, tenant isolation, raw-chunk fields, citations, caches, exports, and sensitive-source leakage in retrieval systems. Use when an authorized RAG assessment must determine what data each identity can retrieve or disclose. - 90multi-tenant-security
Verify tenant isolation in a multi-tenant SaaS — detect IDOR/BOLA, cross-tenant queries, filtering mistakes, leaky caches, jobs, files, search and exports — and enforce tenant scoping by design. Use whenever the app serves several organisations, teams, workspaces or customer accounts. - 85multi-tenant-access-control
Design, implement, or review tenant isolation and authorization across SaaS resources, roles, memberships, workers, caches, and administrative workflows. Use for access decisions or cross-tenant risks; skip authentication-only work and ordinary features with unchanged access boundaries. - 100rate-limit-boundary-fairness
Check that rate limits and quotas are assigned and consumed per identity, so exhaustion of one caller cannot perturb another's authorization guarantees. Use when the API enforces rate limits, quotas or concurrency caps alongside its authorization checks - 100finops-caches
GitHub Actions cache analysis — size, prefix/branch breakdown, stale detection. Use when investigating cache bloat, stale caches, or auditing cache key strategies.