# claxton-anthropic-ai-native-sdlc-playbook-2026-08-21

## Veille

Long-form guide from **Anthropic** by **Louis Claxton** (Applied AI team), published on **August 21, 2026** on the claude.com blog: a stated **40-minute** read, roughly **64,000 characters**, presented as a collection of *plays* drawn from the team's work with its clients. (A) The diagnosis: with code no longer the bottleneck, it shifts to the stages on either side of the build (plan, review/test, deploy), line-by-line controls stop holding once the agent writes most of the diff, and governance cost rises as exceptions still route through periodic committees. (B) The response: six stages (Plan, Design, Build, Test, Deploy, Maintain) organized as a **loop** rather than a chain, each ending with a **committed artifact** that the next stage reads — `intent.md`, `spec.md`, `plan.md`, the diff and its tests, the PR and its findings, the incident record. (1) Institutional knowledge becomes versioned files: `CLAUDE.md`, skills, `REVIEW.md`, `bands.yaml`. (2) Governance splits into two layers, with the skill positioned as an advisory control and the hook as the deterministic layer behind it. Separation of duties is set as an invariant — the agent that writes the code cannot approve it — and the piece closes on *"The loop keeps running. Human judgement stays above it."* The corpus already holds [[clinton-anthropic-secure-ai-native-sdlc-2026-07-21]] on the security side of the same cycle, and [[hingel-augment-how-ai-changes-sdlc-six-stages-2026-06-08]] on the same six-stage breakdown as seen by a competitor.

## Titre Article

The AI-Native SDLC playbook: How to transform your software development lifecycle with AI—stage by stage

## Date

2026-08-21

## URL

https://claude.com/blog/the-ai-native-sdlc-playbook

## Keywords

AI-native SDLC, software development lifecycle, plays, intent.md, spec.md, plan.md, CLAUDE.md, REVIEW.md, bands.yaml, committed artifact, audit trail, plan mode, auto mode, hooks, skills, subagents, parallel sessions, git worktrees, feedback loop, continuous evals, agentic PR review, separation of duties, managed settings, sandbox, MCP, claude-code-action, Agent SDK, control bands, Western Electric rules, OpenTelemetry, DORA, Claude Tag, leading indicators, lagging indicators, governance as code

## Authors

Louis Claxton (Anthropic, équipe Applied AI), sur le blog claude.com ; contributions créditées à Jim Blackhurst, Will Steuk et Jamal Arif.

## Ton

Profile: long-form enterprise guide with an operational aim, "we" voice of a vendor's Applied AI team describing the use of its own products, prescriptive and procedural register, high technical level, aimed at platform leads, tech leads, and compliance/security teams at large organizations, including regulated ones. The structure is that of a **manual** more than an essay: each *play* follows the same grid — what changes, prerequisites, infrastructure, execution steps, governance considerations, a leading indicator and a lagging indicator — and almost every one comes with an artifact shown as-is (`intent.md`, `plan.md`, `CLAUDE.md`, `SKILL.md`, `settings.json`, `bands.yaml`, a GitHub Actions workflow). The vocabulary is borrowed from internal control — *control objectives*, *separation of duties*, *approval gates*, *audit trail*, *blast radius* — and serves to translate agent practices into an auditor's categories. The rhetoric proceeds by systematic binary opposition, each play opening on a *Traditional* / *AI-native* pair. The text owns its commercial role without disguising it: the named products (Claude Code, Claude Design in beta, Claude Tag in public beta, Code Review in *research preview*, Cowork) are its own, and the closing section links to fifteen pages of documentation. It nonetheless remains hedged about its own limits — the skill is said to force nothing, and the managed settings are offered as a starting point to adjust, not a recommendation to copy verbatim.

## Pense-betes

- **Three consequences once the build stops being the constraint**: (1) the bottleneck shifts to the stages that still run at human speed (plan, review/test, deploy); (2) controls become inapplicable — reading every line made sense when a human had written it; (3) governance cost rises, as exceptions route through periodic committees. Example given: a security team sized for human throughput, facing which either the review queue grows or code ships under-reviewed.
- **The throughline is the committed artifact**, not the tool: each stage ends by writing to version control, the next begins by reading it, and the commit chain **is** the audit trail. `.md` dominates upstream because the product owner and the agent read the same file; from the build stage onward, the artifact is the code and its traces.
- **Cascading triggers**: an accepted `intent.md` triggers the requirements/design pass, an approved `spec.md` triggers plan mode, a merged PR triggers the pipeline, a band crossed in production writes the next `intent.md`. Teams start by prompting each stage by hand; the target state is the loop where each accepted artifact arms the next gate.
- **Skill vs. hook — the distinction carries the whole control edifice**: the skill makes policy compliance likely without forcing a session to comply; the hook is deterministic and blocks the action. A policy that must always hold needs a hook or a review pass behind the skill. Corollary: a hook that *requests* human approval belongs in deployment, not in the build, where it would put a person back on the critical path of every parallel session.
- **Legacy systems**: for each artifact, name **one** system as the source of truth (the repo, or Jira/ServiceNow with the `.md` files as working copies), with everything else holding only a link. Simple **chaining** — the artifact carries the record ID, the record carries the commit SHA — is offered as the minimum starting bar.
- **Test**: the feedback loop (tests, build, screenshot diff) runs throughout the task; the verifier subagent is a final pass with fresh context, so the verdict isn't colored by the assumptions that produced the code. For a fix, write the failing test first, commit it, then block the agent from editing it via a hook. Evals are the AI-native counterpart to QA gates: **20 to 50 real tasks** replayed on every change to `CLAUDE.md`, a skill, or a hook, with every incident becoming a permanent eval.
- **Maintain, closing the loop**: **detection stays deterministic** (rolling-window mean and standard deviation, Western Electric rules, a versioned and tested script, no model involved); Claude is invoked only once a band is crossed, and the tier fixes what it can do — 1σ logs, 2σ diagnoses read-only, 3σ proposes (a PR or a pre-approved runbook). Rollback is named as the path that must be the **most rehearsed** in the pipeline.
- ⚠️ **What the piece doesn't quantify**: no quantified results, neither time savings nor adoption rate. The guide's numbers are **implementation parameters** (20-50 evals, two or three parallel sessions to start); outcomes remain metrics to measure yourself, with their source named each time (git log, PR metadata, OpenTelemetry export, DORA, incident-tracking tool).
- **To connect**: [[sfeir-sdlc-ia-cycle-11-phases-2026-06-16]] (a competing breakdown, into eleven phases) and [[sfeir-code-review-anneau-contraintes-2026-07-30]] (the ring of constraints around the agent, where hooks and review are here two distinct rings).

## RésuméDe400mots

Louis Claxton, of Anthropic's Applied AI team, published an implementation guide for an "AI-native" software development lifecycle on August 21, 2026. The starting point is an imbalance: organizations now write code at a speed unimaginable a year earlier, but the processes around it — approval gates, reviews, handoffs, policies — haven't moved. The traditional SDLC was designed for a world where writing code was the longest and costliest stage; its controls also assume that every action is taken by a human.

Three consequences follow. The bottleneck shifts to the stages that still run at human speed, on either side of the build. Controls stop being applicable: reading every line made sense when a person had written it. And governance cost rises, as exceptions route through periodic committees.

The response keeps the control objectives and changes how they're executed. The process becomes a loop, with AI embedded at every point, organized into six stages — Plan, Design, Build, Test, Deploy, Maintain — broken down into *plays* that all follow the same grid, down to the metrics. The throughline is the committed artifact. Intent is captured by its original author as `intent.md`; requirements and design merge into a single session producing `spec.md`, constrained by the brand, security, compliance and UX skills; the build starts in plan mode and locks `plan.md` before any code is written. The commit chain serves as the audit trail.

Institutional knowledge becomes versioned files: `CLAUDE.md` for repository context, skills for cross-cutting policies, `REVIEW.md` for review doctrine, `bands.yaml` for production thresholds. Governance splits into two layers, with the skill as an advisory control and the hook as the deterministic layer that blocks or requests approval. A *managed settings* example details, key by key, what each setting buys in terms of control, from refusing to read secrets to enforcing a minimum version floor.

The Maintain stage closes the loop: a deterministic script monitors a metric, and crossing a band invokes Claude with no human in the call path, at an autonomy level set by the tier. What the agent finds is rewritten as `intent.md` and fed back into the cycle. Claude Tag, in public beta on Slack, extends the pattern to incidents arriving via chat. No quantified results are put forward: the guide provides metrics to measure and names their source.

## GrapheDeConnaissance

- Anthropic —publie→ The AI-Native SDLC playbook (DOCUMENT, 0.97)
- Louis Claxton —a_créé→ The AI-Native SDLC playbook (DOCUMENT, 0.95)
- The AI-Native SDLC playbook —affirme_que→ le goulot se déplace du build vers les étapes restées à vitesse humaine (AFFIRMATION, 0.94)
- SDLC AI-native —est_variante_de→ SDLC (METHODOLOGIE, 0.92)
- SDLC AI-native —utilise→ artefact committé (CONCEPT, 0.93)
- artefact committé —permet→ piste d'audit (CONCEPT, 0.9)
- intent.md —fait_partie_de→ SDLC AI-native (METHODOLOGIE, 0.92)
- spec.md —est_basé_sur→ intent.md (DOCUMENT, 0.91)
- plan.md —est_basé_sur→ spec.md (DOCUMENT, 0.9)
- Plan mode —permet→ plan accepté avant toute écriture de code (CONCEPT, 0.93)
- CLAUDE.md —s_applique_à→ contexte du dépôt lu à chaque session (CONCEPT, 0.92)
- Claude Skills —s_applique_à→ politique appliquée pendant l'écriture du code (CONCEPT, 0.9)
- hooks —améliore→ Claude Skills (TECHNOLOGIE, 0.88)
- Louis Claxton —affirme_que→ une skill est un contrôle consultatif, rien n'oblige une session à la suivre (AFFIRMATION, 0.92)
- REVIEW.md —s_applique_à→ passes de revue bugs, sécurité et conformité au spec et au plan (CONCEPT, 0.89)
- séparation des tâches —s_applique_à→ l'agent qui écrit le code ne peut pas l'approuver (AFFIRMATION, 0.93)
- evals continues —s_applique_à→ configuration d'agent versionnée, testée comme du code (CONCEPT, 0.9)
- Louis Claxton —recommande→ collecter 20 à 50 tâches réelles pour constituer la suite d'evals (AFFIRMATION, 0.88)
- Louis Claxton —recommande→ démarrer à deux ou trois sessions parallèles par ingénieur (AFFIRMATION, 0.87)
- git worktrees —permet→ sessions Claude Code parallèles isolées (CONCEPT, 0.9)
- bands.yaml —permet→ paliers d'autonomie 1σ, 2σ, 3σ (CONCEPT, 0.9)
- détection de bande de contrôle —affirme_que→ la détection reste entièrement déterministe, sans modèle impliqué (AFFIRMATION, 0.92)
- MCP —permet→ déploiement et rollback exposés comme outils cadrés par environnement (CONCEPT, 0.89)
- managed settings —réduit→ surface d'action de l'agent en environnement régulé (CONCEPT, 0.89)
- Claude Tag —s_applique_à→ réponse à incident depuis un canal Slack (CONCEPT, 0.88)
- DORA —mesure→ performance de livraison, indicateur retardé du play CI/CD (MESURE, 0.85)

---
Canonical: https://www.thekb.eu/en/fiches/claxton-anthropic-ai-native-sdlc-playbook-2026-08-21/