← ClaudeAtlas

secure-api-surfacelisted

Product-oriented HTTP API security review (object/function authz, oversharing, mass assignment). Use for public or authenticated APIs, "/api-secure", "BOLA", "BFLA", "IDOR", "API security review", or in Chinese "接口安全", "越权", "水平越权", "垂直越权", "过量数据暴露". Defensive findings only - not red-team exploits or binary audit workflows.
Wanbinyu/api-platform-skills · ★ 1 · AI & Automation · score 69
Install: claude install-skill Wanbinyu/api-platform-skills
# Secure API Surface > UUIDs are not authorization. Specs are not proof. ## Overview Catch API abuse cases product engineers ship by accident. Stay on object-level authz, exposure, mass assignment, injection surfaces, and basic abuse controls. ## Scope | In | Out | |----|-----| | Authn/z on resources | Reverse engineering, exploit PoCs | | Oversharing / DTO leaks | Full org AppSec programs | | Rate limit / pagination caps | Network pentests | | Token/session misuse on APIs | Physical security | ## Steps 1. **Inventory** endpoints (OpenAPI or routes). 2. **Per authenticated resource op** - Object-level (BOLA/IDOR): user A + user B id? - Function-level (BFLA): non-admin -> admin op? - Tenant isolation? 3. **Exposure** - internal fields, tokens, PII on list/detail; chatty errors. 4. **Mass assignment** - client sets `role`, `isAdmin`, `price`, `ownerId`? 5. **Auth gaps** - missing auth; token footguns visible in code. 6. **Injection / SSRF** - query/body/file; user-supplied fetch URLs. 7. **Abuse** - rate limits on auth/expensive routes; max page size; unbounded bulk. 8. **Findings** - severity + concrete fix (code or gateway). ## Exit criteria - [ ] Endpoint inventory - [ ] Object- and function-level checks (or N/A + why) - [ ] Exposure + mass assignment reviewed - [ ] Findings table with severity + fixes - [ ] Residual risk noted - [ ] No exploit payload chains ## Anti-patterns - Generic OWASP essay with no endpoint table - Only finding: "use HTTPS" - Dropp