Enforcement Layer

The enforcement layer is a set of 21 skills loaded at session start that prevent Claude from falling back on training data patterns. These skills enforce correct mental models, forbidden language, forbidden architecture patterns, and self-correction triggers.

Accountability Field Template

schema_version: "1.0"

accountability-field-template.md

Blueprint Routing

**All structured diagram output routes through Blueprint.**

blueprint-routing.md

Concurrent Session Awareness

Chris's default operating method is many simultaneous Claude Code sessions across the cockpit — D drive, monorepo, Main Brain, V2 platform, standalone sites — plus MCP servers running concurrently. Ten concurrent sess...

concurrent-session-awareness.md

Customer Facing Prose Hygiene

VFT's internal AI agents operate with a shared vocabulary that is precise and load-bearing for agent-to-agent work. Codenames, artifact paths, authoring provenance markers, internal process language, and the AI org ch...

customer-facing-prose-hygiene.md

Delegation Manifest Template Leaf Mode

The standard delegation manifest assumes **spawn capability**: an orchestrator (V, Sage, Pax, or a lead command) can spawn specialist agents inline via the Agent tool or Task tool. Specialists receive a bounded work p...

delegation-manifest-template-leaf-mode.md

Ledger Gateway Override

**Every HubSpot write — Listing creation, object updates, association creation, engagement creation, custom event fire, property creation — MUST be routed through Ledger.**

ledger-gateway-override.md

Model Tier Reference

Not every agent needs Opus. The structured prompts, agent definitions, skills, and scripts do the heavy reasoning lifting. The model's job is often "follow these instructions" -- which Sonnet and Haiku handle well. Th...

model-tier-reference.md

Showcase Website Routing

**All public-facing website page implementation routes to Showcase.**

showcase-website-routing.md

UX anti Patterns

Mirror and Showcase reference this during visual QA, portal builds, and interactive audits. Every pattern here is something we've either shipped or could ship if we're not paying attention.

ux-anti-patterns.md

Challenger Coach

This skill governs how agents respond when their conclusions are challenged. It closes the gap between `vf-self-correction.md` (which handles rule violations) and `vf-disagreement-protocol.md` (which handles inter-age...

vf-challenger-coach.md

Conversation Mode

Not every question is a request to do something. Sometimes Chris is thinking out loud, exploring an idea, or asking for analysis — not requesting execution. Misreading conversation mode leads to premature action, wast...

vf-conversation-mode.md

Disagreement Protocol

When AI leaders disagree — V sees one thing, Sage sees another, Pax flags a contradiction — we need a protocol that produces better decisions, not one where whoever speaks last wins.

vf-disagreement-protocol.md

Handoff Protocol

Protocol for work transitions between Claude Desktop (Opus) and Claude Code. Purpose: maintain AI-native framing, prevent translation loss, catch human-paced thinking.

vf-handoff-protocol.md

Operational Verifier

**Before presenting ANY claim about operational state, verify it against real data.**

vf-operational-verifier.md

Output Enforcement

Run this protocol before delivering any document, specification, code, or communication. Purpose: catch violations before output reaches Chris. This protocol applies equally to all AI leaders (V, Sage, Pax) — not just V.

vf-output-enforcement.md

Platform Context

You are operating within the Value-First Platform architecture.

vf-platform-context.md

Preflight Protocol

Run this protocol before implementing any HubSpot-related work. Purpose: surface assumptions, catch bias errors, validate approach against VF Platform architecture.

vf-preflight-protocol.md

Self Correction

Protocol for catching and correcting training data habits without human intervention. The system's job includes catching its own mistakes.

vf-self-correction.md

**If you're about to say "this env var isn't configured" or "I need to find this credential" — READ THIS FILE FIRST.** This file is the authoritative record of what's deployed and where.

vf-tech-stack.md

Validation Gates

Executable verification checkpoints. Run after completing architectural components to verify compliance. Gates pass or fail.

vf-validation-gates.md

Verification Before Completion

No completion claims without fresh verification evidence.

vf-verification-before-completion.md