chore: brand this release v2.0.0

Rename the upcoming release from 0.2.0 to 2.0.0 across docs and the
social-preview card so nothing contradicts: CHANGELOG heading, README roadmap,
brand-packs/README, the-qa-loop, claude-code-workflow, and the hero card
(SocialPreview now renders v2.0.0). The 0.1.0 history and the .archive build
notes are left as written.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

Author: Yuri Strohm

CHANGELOG.md

@@ -2,7 +2,7 @@
 
 All notable changes to shotkit are documented here. Format follows [Keep a Changelog](https://keepachangelog.com/en/1.1.0/), versioning follows [SemVer](https://semver.org/).
 
-## [Unreleased]
+## [2.0.0] - Unreleased
 
 The QA loop closes: the critic now emits a machine-readable verdict, the prompt-forge can act on it, and the capability matrix is guarded so it can't silently rot.
 
@@ -25,7 +25,7 @@ The QA loop closes: the critic now emits a machine-readable verdict, the prompt-
 - Adapter `.md` files now defer to `_capabilities.json` for numeric limits, the JSON owns the numbers, the prose owns the how-to-prompt guidance. Conflicts resolve to the JSON.
 - `install.sh` hardened: array-based execution instead of `eval`, errors to stderr, added `--help`.
 - CI runs the two new validators, including the critique-gate selftest.
-- Regenerated `docs/images/social-preview.png` from a new versioned `SocialPreview` Remotion composition (it was a sourceless baked PNG). The card now reads v0.2.0 and reflects what shipped: `critique.json`, the revise loop (CRITIQUE to FORGE), and brand-lock-extractor feeding the brand-lock bar. The demo.gif and explainer videos are unchanged.
+- Regenerated `docs/images/social-preview.png` from a new versioned `SocialPreview` Remotion composition (it was a sourceless baked PNG). The card now reads v2.0.0 and reflects what shipped: `critique.json`, the revise loop (CRITIQUE to FORGE), and brand-lock-extractor feeding the brand-lock bar. The demo.gif and explainer videos are unchanged.
 
 ### Removed
 
@@ -80,7 +80,7 @@ Initial public release.
 - Works in Claude.ai, Claude Code, Claude API
 - Compatible with the SKILL.md open standard (Codex, Cursor, Gemini CLI, Antigravity, Windsurf, not officially tested)
 
-### Known v0.2.0 work
+### Known v2.0.0 work
 
 - `brand-lock-extractor` skill (PDF/image/URL into brand-lock.md)
 - PDF and PPTX exporters

README.md

@@ -214,7 +214,7 @@ Tested against Claude Opus 4.7 and Claude Sonnet 4.6.
 
 ## Roadmap
 
-v0.2.0 (unreleased) closes the QA loop: structured `critique.json` output, a guarded capability matrix, prompt-forge revision mode, and the fal.ai motion lineup (Kling / Veo / Seedance / Hailuo). See the [changelog](CHANGELOG.md). Still on the roadmap:
+v2.0.0 (unreleased) closes the QA loop: structured `critique.json` output, a guarded capability matrix, prompt-forge revision mode, and the fal.ai motion lineup (Kling / Veo / Seedance / Hailuo). See the [changelog](CHANGELOG.md). Still on the roadmap:
 
 - **`brand-lock-extractor`**. Upload a brand book (PDF, screenshots, URL), get a `brand-lock.md` back. The cold-start killer.
 - **PDF + PPTX exporters**. Siblings to `storyboard-html-preview` for client review and agency handoff.

brand-packs/README.md

@@ -29,7 +29,7 @@ The pack ships with two reference brand packs:
 - **`whystrohm.md`** (flagship). The actual brand pack WhyStrohm uses on its own content. Real palette, real voice rules, real "never" list. Use this as the reference for the level of specificity production work requires.
 - **`examples/saas-clean.md`**. B2B SaaS, restrained, professional. Light backgrounds, single accent. Inter type stack. A neutral counterpoint to the WhyStrohm flagship.
 
-More example brand packs will land in v0.2.0. PRs welcome.
+More example brand packs will land in v2.0.0. PRs welcome.
 
 ## Generating a brand pack from existing assets
 

docs/claude-code-workflow.md

@@ -162,7 +162,7 @@ For day-to-day work, natural-language prompts are sufficient.
 
 ## Updating shotkit
 
-When v0.2.0 ships, update the install:
+When v2.0.0 ships, update the install:
 
 ```bash
 cd shotkit

docs/images/social-preview.png

@@ -35,7 +35,7 @@ The loop runs until every shot is `ACCEPT`, or until you decide a shot is good e
 
 ## What makes it close
 
-Two things the v0.2.0 batch added:
+Two things the v2.0.0 batch added:
 
 1. **A machine-readable verdict.** `visual-asset-critic` writes `output/critique.json` next to the markdown critique. Same review, two surfaces. The markdown is for the human; the JSON is for the loop.
 2. **A prompt-forge that can read it.** `visual-prompt-forge` revision mode takes `shots.json` + one or more `critique.json` files and re-emits prompts for only the shots that failed.

docs/the-qa-loop.md

@@ -19,7 +19,7 @@ const C = {
   muted: '#8C8C92',
 };
 
-const VERSION = 'v0.2.0';
+const VERSION = 'v2.0.0';
 const DATE = '2026-06-18';
 const DOC_ID = 'WS-SK-CARD-001';