All skills
simota avatar

/gateway

@e307415
by shingo imotasimota/agent-skills85 stars
15

Designing and reviewing APIs: OpenAPI spec generation, versioning strategy, breaking change detection, REST/GraphQL best practices. Use for API design or OpenAPI specs.

Use this Skill: https://skilld.dev/gh/simota/agent-skills/gateway

This session only. Nothing lands on disk.

referenceapi-security-anti-patterns.md

≈2k tokens on demand. Your agent reads this file only when SKILL.md points to it.

API Security Anti-Patterns

Failure patterns in authentication/authorization, rate limiting, CORS, data exposure, and the OWASP API Security Top 10

2026-05 baseline: OWASP API Security Top 10 2023 is still the latest edition (no 2025/2026 revision has been released). The newly published OWASP Top 10 for Agentic Applications 2026 (published 2025-12, genai.owasp.org) covers risks on the consuming side of agent/LLM APIs, with ASI01 Agent Goal Hijacking at #1. RFC 9700 / BCP 240 (published 2025-01) formalized as the Best Current Practice for OAuth 2.0 security — mandatory PKCE, deprecation of the Implicit flow, and exact redirect_uri matching are now BCP-grade requirements.

1. The 7 Major API Security Anti-Patterns

# Anti-Pattern Problem Symptoms Countermeasure
AS-01 API Key Only Auth Using an API key as the sole authentication mechanism Key leakage grants full access; rotation is difficult Combine with OAuth 2.0 + mTLS, enforce strict expiry/rotation, IP restrictions/device fingerprinting
AS-02 Broken Object Level Authorization (BOLA) Manipulating object IDs grants access to other users' data #1 on the OWASP API Top 10; BOLA accounts for 40% of all API attacks Authorization checks on every DB operation, unguessable GUIDs, object-level authorization testing
AS-03 Wildcard CORS Access-Control-Allow-Origin: * allows all origins Authenticated requests accepted from any domain Allowlist trusted domains, restrict methods/headers, control credentialed requests
AS-04 No Rate Limiting No request-count limits, vulnerable to DoS/brute force Sudden traffic spikes take down the whole service Request quotas, per-IP limits, exponential backoff, abnormal-traffic alerting
AS-05 Excessive Data Exposure Responses include unnecessary internal data/sensitive info DB structure, internal logs, and personal data leak into responses Return only the minimum necessary data, schema validation, data masking, response auditing
AS-06 Inconsistent Auth Standards Different authentication methods per team/service Ad-hoc workarounds, emerging security gaps Unified cross-service authentication/authorization standard, defined minimum security requirements
AS-07 Poor Logging and Monitoring Auth attempts and anomalous access are not logged Untraceable after an incident, no post-hoc analysis possible Log auth attempts/excessive requests/anomalous patterns, integrate with an observability system

2. OWASP API Security Top 10 (2023)

Rank Risk Overview Detection Pattern
API1 Broken Object Level Authorization Manipulating ID parameters grants access to others' data Tampering the id in GET /api/users/{id} retrieves data
API2 Broken Authentication Flaws in the authentication flow Tokens never expire, weak password policy
API3 Broken Object Property Level Authorization Insufficient authorization at the object property level Mass assignment, returning unnecessary properties
API4 Unrestricted Resource Consumption No limit on resource consumption Unlimited page size, no rate limiting
API5 Broken Function Level Authorization Insufficient authorization at the function level Direct calls to admin APIs, privilege escalation via method changes
API6 Unrestricted Access to Sensitive Business Flows Unrestricted access to sensitive business flows Automated ticket hoarding, mass account creation
API7 Server-Side Request Forgery (SSRF) Server-side request forgery URL input parameters used to access internal services
API8 Security Misconfiguration Security configuration mistakes TLS not configured, debug mode enabled, verbose errors
API9 Improper Inventory Management Poor API asset management Old versions left running, Shadow APIs, undocumented endpoints
API10 Unsafe Consumption of APIs Unsafe consumption of APIs Third-party API responses not validated, redirects followed blindly

3. Authentication/Authorization Pitfalls

Authentication design failures:

  ❌ Bearer Token Without Safeguards:
    → No protection against token leakage, no audience validation, weak session revocation
    → Countermeasure: short expiry, refresh token rotation, strict audience/issuer validation

  ❌ Implicit Flow in Production:
    → Using the OAuth 2.0 Implicit Flow in production
    → Tokens exposed in URL fragments, CSRF vulnerability
    → Countermeasure: use Authorization Code Flow + PKCE (even for SPAs)

  ❌ No Scope Granularity:
    → Only two scopes, `read` / `write`, cover all API access
    → Violates least privilege, grants excessive permissions
    → Countermeasure: fine-grained scopes designed per resource and per operation

  ❌ Missing Token Revocation:
    → Tokens remain valid after a user account is suspended
    → Former employees or compromised accounts retain ongoing access
    → Countermeasure: token blacklisting, short expiry + refresh token rotation

  ❌ Encryption Weakness:
    → Using HTTP, outdated TLS versions, weak cipher algorithms
    → Man-in-the-middle attacks, data interception
    → Countermeasure: enforce TLS 1.2+, AES-256, regular crypto audits

4. Defense-in-Depth Checklist

Defense layers:
  Layer 1 - Gateway: authentication, schema validation, rate limiting, CORS
  Layer 2 - Application: authorization, input validation, business logic verification
  Layer 3 - Data: output filtering, data masking, encryption
  Layer 4 - Network: network segmentation, WAF, DDoS protection
  Layer 5 - Monitoring: logging, anomaly detection, incident response

Key principles:
  - Do not rely on a single defense layer (any layer can be breached)
  - Authentication ≠ authorization (authorization checks are still required after authentication)
  - Validate input on the server side, not just the client side
  - Do not leak internal details in error responses
  - Regular CVE scanning of dependencies

5. Collaboration with Gateway

Usage within Gateway:
  1. SURVEY phase: security screening against AS-01 through AS-07
  2. PLAN phase: cross-check against the OWASP API Top 10
  3. VERIFY phase: confirm authentication/authorization/CORS configuration
  4. PRESENT phase: security improvement report

Quality gates:
  - API-key-only auth → propose OAuth 2.0 + PKCE (prevents AS-01)
  - Direct ID reference with no authorization check → detect BOLA, propose fix (prevents AS-02)
  - CORS `*` setting → propose trusted-domain allowlist (prevents AS-03)
  - Rate limiting not configured → propose 429 + Retry-After implementation (prevents AS-04)
  - Unnecessary fields in responses → design minimal data return (prevents AS-05)
  - Old API versions left running → deprecation + removal plan (prevents OWASP API9)
  - securitySchemes undefined in OpenAPI → add security definitions (prevents misconfiguration)

Source: Nordic APIs: 9 Signs You're Doing API Security Wrong · Levo.ai: REST API Security Best Practices 2026 · OWASP: API Security Top 10 2023 · CyCognito: API Security 2026 Guide · Calmops: API Security Beyond JWT · Wiz: OWASP API Security Top 10

Source: SKILL.md on GitHub

No alerts13d5 checks · Risk SAFE
  • Gen Agent Trust Hub13d

    The skill is a comprehensive API design specialist focusing on OpenAPI, GraphQL, and REST best practices. It provides detailed guidance on versioning, security (OWASP Top 10), and error handling without any malicious code or unsafe operations.

  • Socket13d

    No alerts

  • Snyk13d

    Risk: LOW · No issues

  • Runlayer6mo

    2/14 files flagged

  • ZeroLeaks5mo

    Score: 93/100 · 2 sections analyzed

Signed by skilld at e307415. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 2 days ago.

Activeupdated 2 weeks ago

README badge

README badge for simota/agent-skills/gateway