Standards Overview

Standards Overview

VCP includes 41 standards organized into 12 scopes. All standards are fetched at runtime from the manifest — no local copy is needed.

How Standards Are Loaded

1. Skills (or the shared TypeScript module) fetch manifest.json (v2) from GitHub
2. The v2 manifest's scopes object maps scope names to per-scope manifest files in standards/scopes/
3. Each scope manifest is fetched and its standards are flattened into a unified list
4. Standards are filtered based on the project's .vcp/config.json scopes and compliance settings
5. Each applicable standard is fetched individually using its full HTTPS URL (validated before fetch)
6. The ## Rules section is extracted from each standard

Standards by Scope

Core (12 standards, applies: always)

These standards apply to every VCP-configured project regardless of scope settings.

ID	Severity	Title	Description
core-security	critical	Security	Input validation, output encoding, parameterized queries, secrets management, cryptography, authentication, authorization, deserialization, data protection, timing attacks, key lifecycle, prototype pollution
core-secure-defaults	critical	Secure Defaults	No hardcoded fallback secrets, no default credentials, secure crypto defaults, default-deny permissions, fail-secure, no debug in production, startup validation
core-api-design-security	high	API Design Security	Safe-path-is-easy-path, no caller algorithm selection, semantic types, never ignore security failures, configuration cliffs, explicit opt-in for dangerous ops
core-dependency-management	critical	Dependency Management	Lockfile hygiene, version pinning, package verification, slopsquatting detection, supply chain tools
core-architecture	high	Architecture	Single responsibility, separation of concerns, dependency direction, layer boundaries
core-root-cause-analysis	high	Root Cause Analysis	Fix at the right level, trace to origin, break the death spiral
core-code-quality	high	Code Quality	Consistency, duplication elimination, dead code removal, naming conventions, strategic comments for AI context preservation
core-error-handling	high	Error Handling	Edge cases, boundary validation, structured errors, never swallow errors
core-testing	high	Testing	Test real behavior, meaningful assertions, edge cases, no mock-only tests
core-attack-surface	high	Attack Surface Analysis	Entry point mapping, access level classification, exposure-proportional security, attack surface reduction
core-data-flow-security	critical	Data Flow Security	Source-sink analysis, taint tracking, undefended path detection, ReDoS prevention
core-concurrency-security	high	Concurrency Security	TOCTOU prevention, idempotency keys, database transaction isolation, optimistic concurrency, advisory locks

Web Frontend (4 standards, applies: web-frontend)

ID	Severity	Title	Description
web-frontend-security	critical	Frontend Security	XSS prevention, CSP, CORS, auth token handling, iframe sandboxing, clickjacking, SRI, DOM XSS sinks, postMessage validation, third-party script sandboxing
web-frontend-structure	high	Frontend Structure	Component organization, state management, folder conventions
web-frontend-accessibility	high	Web Accessibility	WCAG 2.2, semantic HTML, keyboard navigation, ARIA, color contrast, focus management
web-frontend-performance	medium	Frontend Performance	Bundle discipline, lazy loading, code splitting, rendering optimization

Web Backend (6 standards, applies: web-backend)

ID	Severity	Title	Description
web-backend-security	critical	Backend Security	Injection prevention, authentication, authorization, secrets management, rate limiting, SSRF, path traversal, file upload, XXE, request smuggling, DB TLS, JWT security, OAuth/OIDC
web-backend-structure	high	Backend Structure	HTTP/business logic separation, service layers, middleware conventions
web-backend-data-access	high	Backend Data Access	Query safety, migration patterns, connection management, ORM usage
web-backend-api-design	high	API Design and Security	REST pagination, RFC 9457 errors, GraphQL depth limiting, gRPC auth, idempotency, persisted queries, unsafe API consumption, business flow protection
web-backend-realtime	high	Realtime Communication	WebSocket auth, origin validation, message schema validation, SSE resumption
web-backend-caching	high	Caching Security	Cache poisoning, cache deception, sensitive data in caches, Redis ACLs

Database (2 standards, applies: database)

ID	Severity	Title	Description
database-encryption	critical	Database Encryption	TDE, column-level encryption, key management, PII protection
database-schema-security	high	Database Schema Security	Row-level security, data classification, audit triggers, masking

Mobile (2 standards, applies: mobile)

ID	Severity	Title	Description
mobile-security	critical	Mobile Security	Keychain/KeyStore, certificate pinning, deep links, biometrics, IPC, background data, app attestation, binary protections, backup exclusion, privacy controls
mobile-platform-configuration	high	Mobile Platform Configuration	ATS, Network Security Config, permissions, WebView, cross-platform frameworks

Desktop (1 standard, applies: desktop)

ID	Severity	Title	Description
desktop-security	critical	Desktop Application Security	Electron context isolation, Tauri capabilities, IPC validation, CSP, auto-update verification, code signing, DevTools, preload integrity, native modules, IPC prototype pollution

CLI (1 standard, applies: cli)

ID	Severity	Title	Description
cli-security-and-quality	high	CLI Security and Quality	Shell injection, argument injection, secure temp files, exit codes, signals, XDG compliance

DevOps (4 standards, applies: devops)

ID	Severity	Title	Description
devops-container-security	critical	Container Security	Non-root containers, multi-stage builds, pinned digests, capability dropping, image scanning, cosign provenance, minimal base images, seccomp/AppArmor, Docker socket, rootless mode
devops-cicd-security	critical	CI/CD Pipeline Security	SHA-pinned actions, minimum permissions, script injection prevention, OIDC auth, SLSA provenance attestation, pull_request_target restrictions
devops-iac-security	high	Infrastructure as Code Security	No hardcoded creds, least-privilege IAM, encrypted storage, locked state files, policy-as-code CI, drift detection, module pinning
devops-kubernetes-security	high	Kubernetes Security	Pod security, RBAC, network policies, external secret managers, mTLS with Istio, secure ingress, Falco runtime monitoring, cluster hardening

Agentic AI (5 standards, applies: agentic-ai)

Standards for AI agent development — MCP servers, multi-agent systems, and LLM tool use. Covers all 10 categories of the OWASP Agentic AI Security Initiative Top 10 (Dec 2025).

ID	Severity	Title	Description
agentic-ai-agent-security	critical	Agent Security	Prompt injection defense (ASI01), sandboxed code execution (ASI05), RAG/memory poisoning prevention (ASI06), human oversight requirements (ASI09)
agentic-ai-tool-security	critical	Tool Security	Tool allowlists, MCP server vetting, tool argument validation, invocation audit logging (ASI02)
agentic-ai-permissions	critical	Agent Permissions	Least privilege, credential scoping, session-scoped tokens, rogue agent detection, filesystem/network isolation (ASI03, ASI10)
agentic-ai-supply-chain	critical	Agentic Supply Chain	MCP server integrity, tool descriptor tampering, model provenance, agent persona verification (ASI04)
agentic-ai-communication	high	Agent Communication	Authenticated inter-agent messaging, message integrity, circuit breakers, cascading failure prevention (ASI07, ASI08)

Compliance (4 standards, applies: compliance:{framework})

These are only loaded when the corresponding compliance framework is configured in .vcp/config.json.

ID	Severity	Applies When	Description
compliance-gdpr	critical	"gdpr" in compliance	Data deletion, retention, consent, PII handling, CCPA/CPRA
compliance-pci-dss	critical	"pci-dss" in compliance	Tokenization, card masking, CDE isolation, PCI DSS v4.0
compliance-hipaa	critical	"hipaa" in compliance	PHI encryption, audit logging, retention, minimum necessary
compliance-accessibility	critical	"accessibility" in compliance	ADA, Section 508/504, EAA, PSBAR, AODA, ACA, EN 301 549, WCAG conformance mapping, accessibility statements, VPAT/ACR, complaint handling

Severity Levels

Severity	Meaning	Examples
critical	Security vulnerabilities or compliance violations	SQL injection, hardcoded secrets, missing encryption
high	Architectural violations that cause long-term damage	SRP violations, layer bleeding, missing error handling
medium	Quality issues that accumulate over time	Code duplication, naming inconsistency, bundle bloat
low	Minor style or convention issues	(None currently — all standards are medium or above)

Cross-References

Standards reference each other and external resources:

• OWASP Top 10:2025 — Web application security risks
• OWASP ASVS v5.0 — Application security verification
• OWASP Mobile Top 10 — Mobile application security risks
• OWASP API Security Top 10 — API security risks
• OWASP Agentic AI Security Top 10 (ASI) — AI agent security risks (ASI01–ASI10)
• CWE (Common Weakness Enumeration) — Vulnerability taxonomy
• WCAG 2.2 — Web accessibility guidelines
• OpenSSF — Open source security guidance

Viewing Standards

All standards are in the standards/ directory with flat {scope}-{topic}.md naming. Scope manifests are in standards/scopes/.

See Standards Format Specification for the structure and writing guidelines.