Replace uses of the legacy `gpt-image-2-all` model with `gpt-image-2` and standardize image workflows: no-reference generation uses POST /v1/images/generations, any-reference flows use POST /v1/images/edits with multipart `image` parts. Update SKILLs, generation scripts, decision logs, and docs to reflect the contract change and edits-vs-generations guidance. Apply corresponding changes across backend (api-server match3d/puzzle modules, openai image adapter, mappers, telemetry, spacetime client/module), frontend components and services (Match3D, Puzzle, CreativeImageInputPanel, runtime shells), and add new spritesheet/parser files and tests. Also add media/logo.png. These changes align repository code and documentation with the VectorEngine image API contract and update generation/upload handling (green-screen -> alpha processing, spritesheet handling, and related tests).
Rewrite Genarrative play-type integration guidance across .codex and .hermes to define a platform-level SOP: default to form/image workbench, unify single-image asset slots (CreativeImageInputPanel), standardize series-material sheet->cut->transparent->OSS pipeline, and forbid copying legacy chat/agent workflows as the default. Add decision-log entry freezing the SOP and a pitfalls note warning against direct reuse of old play tools. Update CONTEXT.md and docs/README.md, add a new PRD file, and apply related small server-side changes (module-auth, spacetime-client mappers and runtime) to align back-end code with the new contracts and flows.
Allow history-generated image paths to be submitted where Data URLs were previously required and avoid treating partial/result-page generations as blocking the whole draft. Backend: resolve history /generated-* references via resolve_puzzle_reference_image_as_data_url and convert to PuzzleDownloadedImage; add PuzzleDownloadedImage::from_resolved_reference_image; extend draft handling to apply generated level metadata (auto-naming) and normalize generation_status to treat levels with images as ready. API: add shouldAutoNameLevel to action contracts and use it to request/refine generated level names. Spacetime/module and mappers: normalize completed level statuses when saving/reading so result-page background or per-level generation doesn't mask completed drafts. Frontend: expose resolver helpers, only mark a work as generating when no usable cover or ready level exists, keep level controls enabled during UI-background regeneration, and add tests covering history-image submission, auto-naming, and UI-background/partial-generation behaviors.
Allow local env files to reliably override authentication feature flags (SMS/WeChat) by whitelisting keys in scripts/dev-utils.mjs and adding a unit test. Add SMS checks to scripts/check-api-server-env.mjs. Make server config.parse_bool tolerant of shell-wrapped quoted values (e.g. '"true"') and add tests so SMS_AUTH_ENABLED is parsed correctly when shells supply quotes. Update docs to clarify SMS env behaviour, restart requirements, and add guidance + a CSS fallback for old mobile browsers (QQ/X5) so public cover images render even when aspect-ratio is unsupported. Also include related frontend test and component adjustments and add puzzle onboarding handlers/endpoints in server-rs/crates/api-server/src/puzzle.rs.
