ray-amjad-agentic-coding-ladder

name	Ray Amjad — Agentic Coding Ladder
description	The implementation layer of Boris Cherny's 4-level AI-adoption ladder: level diagnosis, task-lifecycle forging, verification environments, loops/routines, fan-out missions, chat-native harnesses. Trust = artifacts, never claims.
version	1.0
workflows	9
source	YouTube XLA-sTSJ-Wc (2026-07-17) — watched via /watch, transcript + 48 frames

Ray Amjad — Agentic Coding Ladder

Boris Cherny's ladder describes the stages of agentic-coding adoption; Ray Amjad demonstrates the buildable mechanisms for climbing it — with receipts. The unit of trust is the artifact (recording / PR / MP3), never the agent's claim. Load genius.md before any workflow.

Available Workflows

Tier 1 — Foundation

Workflow	Produces	Use When
ladder-diagnostic	Level diagnostic card (actual vs perceived + missing mechanism)	Anyone claims (or wonders) what level they're at
task-lifecycle-forge	Paste-ready Task Lifecycle file (CLAUDE.md block or skill)	A repo needs spec-in → recording+PR-out autonomy
verification-environment	Provisioning plan (surface→instrument→artifact table + build order)	Verification is shallow; agent claims can't be trusted end-to-end
level-up-plan	One-rung climb plan with trust exit-test	Diagnosis done; time to climb exactly one level

Tier 2 — Practitioner

Workflow	Produces	Use When
loop-hunter	Routine sheet (evidence-backed loop specs + meta-loop)	Recurring manual work suspected; the 2→3 climb
fanout-mission	Coordinator fan-out mission prompt (7-part grammar)	A queue of similar tasks + a trusted per-item lifecycle
chat-native-harness	Harness blueprint (runtime, skills, channels, grammar)	Moving lifecycles off the laptop into Slack-native ops

Tier 3 — Stacking

Workflow	Produces	Use When	Stacks With
system-ladder-audit	System audit card (mechanics table + top-5 upgrades)	Auditing a built Claude Code system, incl. Antigravity	boris-claude-code, orchestration-doctrine
adoption-brief	≤2-page client-facing adoption brief	Consulting/diagnostic engagements	geoff-woods-ai-thought-partner, Proof-to-Market offers

Stacking Guide

• boris-claude-code — philosophy ↔ execution: Boris's plan-first/self-verifying concepts ground every workflow here; run his agentic-product-engineering-validation for the product frame, then task-lifecycle-forge for the mechanism.
• nick-saraev-agentic-workflows — bottleneck thinking feeds loop-hunter evidence harvesting.
• nate-b-jones-trust-architecture — deepens the trust-ledger phase of level-up-plan.
• Wargame OS — failure maps ARE the trust ledger; bank them before widening autonomy.
• geoff-woods-ai-thought-partner — wraps adoption-brief for non-technical operators.

Quick Reference

• Genius Context: genius.md — 12 patterns, hidden knowledge, exemplars, rubric
• Source ladder: references/boris-ladder-source.md — Boris's table, verbatim; quote, never extend
• Corpus anchors: extractions/ray-amjad/reference-corpus/ — his real Task Lifecycle file + Sentry fan-out prompt
• The two-test challenge (fastest diagnostic): trusted end-to-end self-verification? automated review with severity routing? Both no → Level 1, whatever it looks like.

Execution Prompts (structure-pure v2)

9 deterministic practitioner prompts — each carries an Output Contract, Output Skeleton, and Quality Gate. When a deliverable matches one, Read it and honor its contract instead of improvising the output shape.

• Ray Amjad — Client Adoption Brief — skills/ray-amjad-agentic-ladder/references/prompts-v2/adoption-brief.md
• Ray Amjad — Chat-Native Harness Blueprint — skills/ray-amjad-agentic-ladder/references/prompts-v2/chat-native-harness.md
• Ray Amjad — Fan-Out Mission Prompt — skills/ray-amjad-agentic-ladder/references/prompts-v2/fanout-mission.md
• Ray Amjad — Ladder Diagnostic Card — skills/ray-amjad-agentic-ladder/references/prompts-v2/ladder-diagnostic.md
• Ray Amjad — Level-Up Plan (One Rung) — skills/ray-amjad-agentic-ladder/references/prompts-v2/level-up-plan.md
• Ray Amjad — Routine Sheet (Loop Hunter) — skills/ray-amjad-agentic-ladder/references/prompts-v2/loop-hunter.md
• Ray Amjad — System Ladder Audit Card — skills/ray-amjad-agentic-ladder/references/prompts-v2/system-ladder-audit.md
• Ray Amjad — Task Lifecycle File — skills/ray-amjad-agentic-ladder/references/prompts-v2/task-lifecycle-forge.md
• Ray Amjad — Verification Environment Plan — skills/ray-amjad-agentic-ladder/references/prompts-v2/verification-environment.md