Skill · Portfolio

디자인 포트폴리오 구성·첨삭

리서치·스케치·렌더·시제품 자료로 근거 있는 가로형 포트폴리오를 구성하고 첨삭하는 원작 스킬(산업디자인). 원작자 truman-t3, MIT.

0installs
SkillIncludes scriptsApprovedBy truman-t3 · MIT · Uploaded by feeeldstudy · Version 1Claude · ChatGPT · Codex · Cursor
industrial-design-portfolio/showcase/modular-desk-lamp/source_notes.md
# Source NotesThis is a fictional demonstration project created to show how the Skill structures an industrial design case study. It does not document a completed design project.- No interviews, observations, participants, prototypes, measurements, costs, supplier reviews, patents, or launched outcomes are claimed.- Behavioral scenarios are synthetic and classified as E0.- Product architecture and manufacturing statements are design intent, not engineering proof.- Every generated visual is labeled next to the image in the finished portfolio.- All unresolved claims remain in the manifest with `caveat` or `to_validate` status.## Public reproduction chainThe repository includes the materials needed to inspect or repeat the editorial workflow:- `input/brief.md` — the only intake artifact;- `input/asset-inventory.json` — separates intake material, later generated visuals, and missing evidence;- `input/gating-questions.md` — questions the Agent should resolve before planning pages;- `input/reproduction-prompt.md` — a reusable forward-test prompt, not an original transcript;- `input-to-output-trace.md` — maps inputs and created artifacts to the finished pages.These files document a fictional reference case. They do not replace the separate sanitized real-project forward test still required to demonstrate performance on messy, authentic project material.## Generated asset registerAll seven source PNG images were generated with the built-in image-generation capability. The final Hero established the reference language: circular weighted base, two slim arms, cylindrical joints, capsule head, deep-blue matte aluminum, warm diffuser, and cobalt replaceable sleeve. Other images were then generated to preserve that language. Browser-facing WebP derivatives are created at quality 82 with `showcase/tools/optimize_images.py`. The editable source PNG files are retained locally for future editing but intentionally excluded from the lightweight Git repository.| File | Purpose | Evidence boundary ||---|---|---|| `images/generated/lamp-12-final-hero.webp` | Product-language master and final Hero | Concept visualization only || `images/generated/lamp-04-scenarios-sheet.webp` | Task, video-call, and compact-desk scenarios | Synthetic; not research evidence || `images/generated/lamp-06-concepts-sheet.webp` | Clamp, weighted-base, and rail routes | Fictional concept divergence || `images/generated/lamp-08-cmf.webp` | Form and CMF detail | Intent; not material approval || `images/generated/lamp-09-conceptual-architecture.webp` | Exploded module relationship | Conceptual architecture illustration; not engineering evidence || `images/generated/lamp-10-validation-plan.webp` | Reach, stability, joint-effort, and glare questions | Future test plan; no prototype or result || `images/generated/lamp-11-iteration-pair.webp` | Proposed stability revision | Illustrative comparison; no measured improvement |The visuals intentionally contain no brand, dimensions, engineering callouts, claimed measurements, participant imagery, or test results.