[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M78FEK3EGM73FSXCESK48JRB":3,"skill-same:01M78FEK3EGM73FSXCESK48JRB":50},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":17,"publishedAt":18,"linkOnly":19,"markdownContent":20,"markdownBody":21,"frontMatter":22,"refName":23,"commitSha":24,"skillDirectory":25,"forksCount":26,"repositoryHtmlUrl":27,"repositoryHomepageUrl":6,"ownerHtmlUrl":28,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":29,"install":30,"resources":43},"01M78FEK3EGM73FSXCESK48JRB","run-hncs",null,"Run, build, test, smoke-test or screenshot the Hncs camera color-science project - apply a brand look to a photo, render a contact sheet of every shipped look, run the CLI tools (LUT export, RAW→Log, video, brand classifier), or check which entry points work in this checkout. Use when asked to run Hncs, render\u002Fpreview a brand look, verify apply_* output, or reproduce a research experiment.","songjiun10-collab\u002FHncs","songjiun10-collab","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F293141730?v=4","01M78FEF4NYSK5BRK9HBBT64X1","01M78FEF5AKPHSAAMYT6249TR7",16,"MIT","https:\u002F\u002Fgithub.com\u002Fsongjiun10-collab\u002FHncs\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Frun-hncs\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: run-hncs\ndescription: Run, build, test, smoke-test or screenshot the Hncs camera color-science project - apply a brand look to a photo, render a contact sheet of every shipped look, run the CLI tools (LUT export, RAW→Log, video, brand classifier), or check which entry points work in this checkout. Use when asked to run Hncs, render\u002Fpreview a brand look, verify apply_* output, or reproduce a research experiment.\n---\n\n# Running Hncs\n\nPython image-in\u002Fimage-out library (`brands\u002F*.py` `apply_*` functions) plus\n9 user-facing CLIs under `tools\u002Fcli\u002F` (plus ~140 research scripts elsewhere\nunder `tools\u002F`). No server. There is a Tkinter GUI (`python3 -m gui`), but it\nneeds a display and a Python built with `tkinter`, so it isn't the agent\npath — the equivalent of a screenshot here is **a rendered output image you\nactually open and look at**.\n\nDriven by `.claude\u002Fskills\u002Frun-hncs\u002Fdriver.py`. All paths below are\nrelative to the repo root, and everything must run from the repo root\n(the `core`\u002F`brands`\u002F`tools` imports are path-relative).\n\n## Prerequisites\n\nAlready present in the standard container. On a bare machine:\n\n```bash\npip install -r requirements.txt\napt-get install -y libimage-exiftool-perl   # hybrid_engine + contributed-data verification\n```\n\n`darktable-cli` (`apt-get install -y darktable`, ~180 packages) is needed\n**only** to reproduce `tools\u002Fresearch\u002Fevaluate_darktable_vs_rawpy.py`. Nothing else\nuses it.\n\n## Check what this checkout can actually run\n\nMost interesting entry points need image corpora that are **git-ignored**,\nso a fresh clone can run far less than the README implies. Start here:\n\n```bash\npython3 .claude\u002Fskills\u002Frun-hncs\u002Fdriver.py env\n```\n\nPrints dependency status, external binaries, and dataset presence:\n\n```\n== 데이터셋 (전부 .gitignore - 신선한 클론엔 없음) ==\n  OK    raw_calib_cache\u002F           30개  - 핫셀블라드 raw+jpeg 13쌍 - calibrate \u002F evaluate_* \u002F hybrid_engine.main\n  OK    downloaded_samples\u002F       123개  - population 샘플 - tools.cli.analyze (없으면 네트워크에서 받음)\n  OK    raw_calib_cache_fuji\u002F      10개  - 후지 raw+jpeg 3쌍 - evaluate_fuji_demosaic\n```\n\n## Run (agent path)\n\n### Smoke — everything that works with no external data\n\n```bash\npython3 .claude\u002Fskills\u002Frun-hncs\u002Fdriver.py smoke\n```\n\nExercises the library entry point plus 5 CLIs, exits non-zero on any\nfailure. Takes ~1 min (the `denoise` step dominates). Verified output:\n\n```\n  PASS  라이브러리: apply_hncs 직접 호출\n  PASS  CLI: tools.cli.export_lut --list\n  PASS  CLI: tools.cli.export_lut (17격자 .cube 생성)\n  PASS  CLI: tools.cli.classify_brand predict\n  PASS  CLI: tools.cli.denoise\n  PASS  CLI: hybrid_engine.convert (--source 필수)\n\n6 pass \u002F 0 fail\n```\n\n### Contact sheet — the visual verification\n\n```bash\npython3 .claude\u002Fskills\u002Frun-hncs\u002Fdriver.py sheet\n# -> \u002Ftmp\u002Fhncs_run\u002Fcontact_sheet.png   (57 tiles: original + 56 looks, 2026-09-25)\n```\n\n**Then open the PNG and look at it.** Distinct looks should be obvious:\n`eterna_cinema` teal, `nostalgic_neg` magenta-shifted, `acros`\u002F\n`monochrome` black-and-white, `hncs` brightened warm, the ten\npopulation-fit brands subtler. Uniform tiles = the pipeline is broken.\n\nThis is the one command that proves the whole rendering path end-to-end.\n\n### One look at full resolution\n\n```bash\npython3 .claude\u002Fskills\u002Frun-hncs\u002Fdriver.py look hasselblad\npython3 .claude\u002Fskills\u002Frun-hncs\u002Fdriver.py look astia --src photo.jpg --out \u002Ftmp\u002Fout.png\n```\n\nAccepts a bare brand (`hasselblad`), a preset (`astia`), or the full\nfunction name (`apply_canon_look`). Unknown name prints the full list.\nWith no `--src` it crops a clean frame out of a committed demo image, so\nit works with no data at all.\n\nOutput dir defaults to `\u002Ftmp\u002Fhncs_run`; override with `HNCS_DRIVER_OUT`.\n\n### Direct invocation (what most PRs touch)\n\nChanges here are usually to a single `apply_*` or a `core\u002F` helper. Skip\nthe CLI entirely:\n\n```bash\npython3 -c \"\nimport cv2\nfrom brands.hasselblad import apply_hncs\nsrc = cv2.imread('docs\u002Fimages\u002Fbefore_after_hncs.jpg')[60:, :903]\ncv2.imwrite('\u002Ftmp\u002Fout.png', apply_hncs(src))\n\"\n```\n\n### Tests\n\n```bash\npython3 -m unittest discover -s tests     # 1442 tests (2026-09-25), ~40s, no external data needed\n```\n\n## Run (CLI tools individually)\n\nAll verified working in this container:\n\n```bash\npython3 -m tools.cli.export_lut --list\npython3 -m tools.cli.export_lut hasselblad \u002Ftmp\u002Fhb.cube --size 17\npython3 -m tools.cli.classify_brand predict \u002Ftmp\u002Fhncs_run\u002F_smoke_src.jpg\npython3 -m tools.cli.denoise in.jpg out.jpg --strength 5\npython3 -m tools.cli.video_engine in.mp4 out.mp4 --brand hasselblad\npython3 -m hybrid_engine.convert in.jpg out.jpg --source nikon --target hasselblad\npython3 -m tools.cli.raw_pipeline raw_calib_cache\u002F00378.jpg.3FR \u002Ftmp\u002Fo.tiff --log-space V-Log\n```\n\nThe last one needs a RAW file — any `.3FR`\u002F`.RAF`\u002F`.CR3`\u002F`.ARW` works, but\n`raw_calib_cache\u002F` is git-ignored, so supply your own on a fresh clone.\n\n## Gotchas\n\n- **`hybrid_engine.convert` hard-fails on an image with no EXIF** —\n  `EXIF로 소스 브랜드를 못 알아봄 (Make=None, Model=None)`, exit 1. Any\n  image written by `cv2.imwrite` has no EXIF, so anything you generate\n  mid-pipeline hits this. Pass `--source nikon` (or whichever brand).\n- **`tools.cli.lens_correction` hard-fails the same way** —\n  `--focal-length가 없고 EXIF에 FocalLength도 없음`. Pass\n  `--focal-length` and `--aperture` explicitly.\n- **`tools.cli.analyze \u003Cbrand>` is not read-only.** It downloads the sample\n  corpus (139 images for `hasselblad`) into git-ignored\n  `downloaded_samples\u002F` on first run — it only skips what's already\n  cached (`os.path.exists`). Needs network plus the allowlisted domains in\n  `.claude\u002Fsettings.json`. Don't run it casually to \"see what it does.\"\n- **`apply_acros` and `apply_monochrome` return 2D single-channel**, not\n  3-channel BGR — deliberate, pinned by\n  `test_mono_presets_return_single_channel`, and why `export_lut` refuses\n  them (`알 수 없는 preset: fuji_acros`). Anything tiling `apply_*` output\n  must `cv2.cvtColor(out, cv2.COLOR_GRAY2BGR)` first. The driver does.\n- **`docs\u002Fimages\u002F*.jpg` are composites with baked-in labels**, not clean\n  source photos. The top ~60px is a label bar and `before_after_hncs.jpg`\n  is a side-by-side pair. Crop `[60:, :width\u002F\u002F2]` for a usable frame —\n  `driver.py`'s `default_source()` does exactly this.\n- **`colour-science` prints a matplotlib warning on every import.**\n  Harmless noise, appears in front of real output; filter it when reading\n  logs.\n- **Research scripts (`tools\u002Fevaluate_*.py`) take hours**, not minutes —\n  the chromatic-aberration one ran ~2h on 13 RAW pairs. Background them\n  and watch the log; see `tools\u002FCLAUDE.md`.\n\n## Troubleshooting\n\n| Symptom | Fix |\n|---|---|\n| `ModuleNotFoundError: No module named 'core'` | You're not in the repo root. `cd` there. |\n| `EXIF로 소스 브랜드를 못 알아봄` | Add `--source \u003Cbrand>` to `hybrid_engine.convert`. |\n| `--focal-length가 없고 EXIF에 FocalLength도 없음` | Add `--focal-length N --aperture N` to `tools.cli.lens_correction`. |\n| `알 수 없는 preset: fuji_acros` from `export_lut` | Expected — monochrome presets are excluded from LUT export. Use `--list`. |\n| `could not broadcast input array from shape (H,W) into (H,W,3)` | You got a mono preset's 2D output. `cv2.cvtColor(out, cv2.COLOR_GRAY2BGR)`. |\n| `FileNotFoundError` under `raw_calib_cache\u002F` | Git-ignored dataset absent. Run `driver.py env` to confirm, then supply your own RAW+JPEG pairs. |\n| Sheet tiles all look identical | Real breakage in the shared curve\u002Fengine code — check `core\u002Fcurve.py` and `core\u002Fengine.py`. |\n","# Running Hncs\n\nPython image-in\u002Fimage-out library (`brands\u002F*.py` `apply_*` functions) plus\n9 user-facing CLIs under `tools\u002Fcli\u002F` (plus ~140 research scripts elsewhere\nunder `tools\u002F`). No server. There is a Tkinter GUI (`python3 -m gui`), but it\nneeds a display and a Python built with `tkinter`, so it isn't the agent\npath — the equivalent of a screenshot here is **a rendered output image you\nactually open and look at**.\n\nDriven by `.claude\u002Fskills\u002Frun-hncs\u002Fdriver.py`. All paths below are\nrelative to the repo root, and everything must run from the repo root\n(the `core`\u002F`brands`\u002F`tools` imports are path-relative).\n\n## Prerequisites\n\nAlready present in the standard container. On a bare machine:\n\n```bash\npip install -r requirements.txt\napt-get install -y libimage-exiftool-perl   # hybrid_engine + contributed-data verification\n```\n\n`darktable-cli` (`apt-get install -y darktable`, ~180 packages) is needed\n**only** to reproduce `tools\u002Fresearch\u002Fevaluate_darktable_vs_rawpy.py`. Nothing else\nuses it.\n\n## Check what this checkout can actually run\n\nMost interesting entry points need image corpora that are **git-ignored**,\nso a fresh clone can run far less than the README implies. Start here:\n\n```bash\npython3 .claude\u002Fskills\u002Frun-hncs\u002Fdriver.py env\n```\n\nPrints dependency status, external binaries, and dataset presence:\n\n```\n== 데이터셋 (전부 .gitignore - 신선한 클론엔 없음) ==\n  OK    raw_calib_cache\u002F           30개  - 핫셀블라드 raw+jpeg 13쌍 - calibrate \u002F evaluate_* \u002F hybrid_engine.main\n  OK    downloaded_samples\u002F       123개  - population 샘플 - tools.cli.analyze (없으면 네트워크에서 받음)\n  OK    raw_calib_cache_fuji\u002F      10개  - 후지 raw+jpeg 3쌍 - evaluate_fuji_demosaic\n```\n\n## Run (agent path)\n\n### Smoke — everything that works with no external data\n\n```bash\npython3 .claude\u002Fskills\u002Frun-hncs\u002Fdriver.py smoke\n```\n\nExercises the library entry point plus 5 CLIs, exits non-zero on any\nfailure. Takes ~1 min (the `denoise` step dominates). Verified output:\n\n```\n  PASS  라이브러리: apply_hncs 직접 호출\n  PASS  CLI: tools.cli.export_lut --list\n  PASS  CLI: tools.cli.export_lut (17격자 .cube 생성)\n  PASS  CLI: tools.cli.classify_brand predict\n  PASS  CLI: tools.cli.denoise\n  PASS  CLI: hybrid_engine.convert (--source 필수)\n\n6 pass \u002F 0 fail\n```\n\n### Contact sheet — the visual verification\n\n```bash\npython3 .claude\u002Fskills\u002Frun-hncs\u002Fdriver.py sheet\n# -> \u002Ftmp\u002Fhncs_run\u002Fcontact_sheet.png   (57 tiles: original + 56 looks, 2026-09-25)\n```\n\n**Then open the PNG and look at it.** Distinct looks should be obvious:\n`eterna_cinema` teal, `nostalgic_neg` magenta-shifted, `acros`\u002F\n`monochrome` black-and-white, `hncs` brightened warm, the ten\npopulation-fit brands subtler. Uniform tiles = the pipeline is broken.\n\nThis is the one command that proves the whole rendering path end-to-end.\n\n### One look at full resolution\n\n```bash\npython3 .claude\u002Fskills\u002Frun-hncs\u002Fdriver.py look hasselblad\npython3 .claude\u002Fskills\u002Frun-hncs\u002Fdriver.py look astia --src photo.jpg --out \u002Ftmp\u002Fout.png\n```\n\nAccepts a bare brand (`hasselblad`), a preset (`astia`), or the full\nfunction name (`apply_canon_look`). Unknown name prints the full list.\nWith no `--src` it crops a clean frame out of a committed demo image, so\nit works with no data at all.\n\nOutput dir defaults to `\u002Ftmp\u002Fhncs_run`; override with `HNCS_DRIVER_OUT`.\n\n### Direct invocation (what most PRs touch)\n\nChanges here are usually to a single `apply_*` or a `core\u002F` helper. Skip\nthe CLI entirely:\n\n```bash\npython3 -c \"\nimport cv2\nfrom brands.hasselblad import apply_hncs\nsrc = cv2.imread('docs\u002Fimages\u002Fbefore_after_hncs.jpg')[60:, :903]\ncv2.imwrite('\u002Ftmp\u002Fout.png', apply_hncs(src))\n\"\n```\n\n### Tests\n\n```bash\npython3 -m unittest discover -s tests     # 1442 tests (2026-09-25), ~40s, no external data needed\n```\n\n## Run (CLI tools individually)\n\nAll verified working in this container:\n\n```bash\npython3 -m tools.cli.export_lut --list\npython3 -m tools.cli.export_lut hasselblad \u002Ftmp\u002Fhb.cube --size 17\npython3 -m tools.cli.classify_brand predict \u002Ftmp\u002Fhncs_run\u002F_smoke_src.jpg\npython3 -m tools.cli.denoise in.jpg out.jpg --strength 5\npython3 -m tools.cli.video_engine in.mp4 out.mp4 --brand hasselblad\npython3 -m hybrid_engine.convert in.jpg out.jpg --source nikon --target hasselblad\npython3 -m tools.cli.raw_pipeline raw_calib_cache\u002F00378.jpg.3FR \u002Ftmp\u002Fo.tiff --log-space V-Log\n```\n\nThe last one needs a RAW file — any `.3FR`\u002F`.RAF`\u002F`.CR3`\u002F`.ARW` works, but\n`raw_calib_cache\u002F` is git-ignored, so supply your own on a fresh clone.\n\n## Gotchas\n\n- **`hybrid_engine.convert` hard-fails on an image with no EXIF** —\n  `EXIF로 소스 브랜드를 못 알아봄 (Make=None, Model=None)`, exit 1. Any\n  image written by `cv2.imwrite` has no EXIF, so anything you generate\n  mid-pipeline hits this. Pass `--source nikon` (or whichever brand).\n- **`tools.cli.lens_correction` hard-fails the same way** —\n  `--focal-length가 없고 EXIF에 FocalLength도 없음`. Pass\n  `--focal-length` and `--aperture` explicitly.\n- **`tools.cli.analyze \u003Cbrand>` is not read-only.** It downloads the sample\n  corpus (139 images for `hasselblad`) into git-ignored\n  `downloaded_samples\u002F` on first run — it only skips what's already\n  cached (`os.path.exists`). Needs network plus the allowlisted domains in\n  `.claude\u002Fsettings.json`. Don't run it casually to \"see what it does.\"\n- **`apply_acros` and `apply_monochrome` return 2D single-channel**, not\n  3-channel BGR — deliberate, pinned by\n  `test_mono_presets_return_single_channel`, and why `export_lut` refuses\n  them (`알 수 없는 preset: fuji_acros`). Anything tiling `apply_*` output\n  must `cv2.cvtColor(out, cv2.COLOR_GRAY2BGR)` first. The driver does.\n- **`docs\u002Fimages\u002F*.jpg` are composites with baked-in labels**, not clean\n  source photos. The top ~60px is a label bar and `before_after_hncs.jpg`\n  is a side-by-side pair. Crop `[60:, :width\u002F\u002F2]` for a usable frame —\n  `driver.py`'s `default_source()` does exactly this.\n- **`colour-science` prints a matplotlib warning on every import.**\n  Harmless noise, appears in front of real output; filter it when reading\n  logs.\n- **Research scripts (`tools\u002Fevaluate_*.py`) take hours**, not minutes —\n  the chromatic-aberration one ran ~2h on 13 RAW pairs. Background them\n  and watch the log; see `tools\u002FCLAUDE.md`.\n\n## Troubleshooting\n\n| Symptom | Fix |\n|---|---|\n| `ModuleNotFoundError: No module named 'core'` | You're not in the repo root. `cd` there. |\n| `EXIF로 소스 브랜드를 못 알아봄` | Add `--source \u003Cbrand>` to `hybrid_engine.convert`. |\n| `--focal-length가 없고 EXIF에 FocalLength도 없음` | Add `--focal-length N --aperture N` to `tools.cli.lens_correction`. |\n| `알 수 없는 preset: fuji_acros` from `export_lut` | Expected — monochrome presets are excluded from LUT export. Use `--list`. |\n| `could not broadcast input array from shape (H,W) into (H,W,3)` | You got a mono preset's 2D output. `cv2.cvtColor(out, cv2.COLOR_GRAY2BGR)`. |\n| `FileNotFoundError` under `raw_calib_cache\u002F` | Git-ignored dataset absent. Run `driver.py env` to confirm, then supply your own RAW+JPEG pairs. |\n| Sheet tiles all look identical | Real breakage in the shared curve\u002Fengine code — check `core\u002Fcurve.py` and `core\u002Fengine.py`. |\n",{"name":5,"description":7},"main","513567c2b0eb44c159ebac74b0491bdcf823c377",".claude\u002Fskills\u002Frun-hncs",2,"https:\u002F\u002Fgithub.com\u002Fsongjiun10-collab\u002FHncs","https:\u002F\u002Fgithub.com\u002Fsongjiun10-collab","2026-11-07 16:34:56",{"source":8,"skillName":5,"commands":31,"prompt":41,"zipAvailable":42},[32,35,38],{"tool":33,"command":34},"npx","npx skills add songjiun10-collab\u002FHncs --skill run-hncs",{"tool":36,"command":37},"bunx","bunx skills add songjiun10-collab\u002FHncs --skill run-hncs",{"tool":39,"command":40},"pnpm","pnpm dlx skills add songjiun10-collab\u002FHncs --skill run-hncs","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fsongjiun10-collab\u002FHncs\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Frun-hncs\u002FSKILL.md",true,[44],{"path":45,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":48,"sourceHtmlUrl":49},"driver.py","SCRIPT","脚本",12174,"https:\u002F\u002Fgithub.com\u002Fsongjiun10-collab\u002FHncs\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Frun-hncs\u002Fdriver.py",[51,58,65,72,79,86],{"skillNo":52,"name":53,"nameZh":6,"description":54,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":55,"categories":56,"keywords":57,"publishedAt":18},"01M78FJ7DGE8S2MHJQY4RVTPMF","widen-the-solution-space","Deliberately widen the set of candidate approaches before committing to one. Use before irreversible decisions (architecture, DB schema, data format, public API), when the first solution feels complicated or awkward, when the user is deciding rather than instructing (\"what's the best way to do this?\"), or when a previous attempt at the same problem failed. Skip for routine work with one obvious answer.","https:\u002F\u002Fgithub.com\u002Fsongjiun10-collab\u002FHncs\u002Fblob\u002Fmain\u002Fsenior-engineer-mindset\u002Fwiden-the-solution-space\u002FSKILL.md",[],[],{"skillNo":59,"name":60,"nameZh":6,"description":61,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":62,"categories":63,"keywords":64,"publishedAt":18},"01M78FJ2D6HMAGS8E11VBD4NS0","weigh-tradeoffs","Compare real alternatives explicitly and record why one was chosen. Use when picking between libraries, data structures, architectures, or storage options; when a decision is hard to reverse (schema, public API, data format, dependency); or when the user asks \"which is better, A or B\". Also use to decide how much deliberation a decision even deserves.","https:\u002F\u002Fgithub.com\u002Fsongjiun10-collab\u002FHncs\u002Fblob\u002Fmain\u002Fsenior-engineer-mindset\u002Fweigh-tradeoffs\u002FSKILL.md",[],[],{"skillNo":66,"name":67,"nameZh":6,"description":68,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":69,"categories":70,"keywords":71,"publishedAt":18},"01M78FHX1P3QMTMC8WPQS0Z669","verify-before-claiming","Run the verification and read its output before saying anything is done, fixed, working, or passing. Use before any completion claim, before committing or opening a PR, before moving to the next task, and after a subagent or tool reports success. Triggers on the words \"should work\", \"fixed\", \"done\", \"passing\", \"looks good\" — and on the urge to celebrate.","https:\u002F\u002Fgithub.com\u002Fsongjiun10-collab\u002FHncs\u002Fblob\u002Fmain\u002Fsenior-engineer-mindset\u002Fverify-before-claiming\u002FSKILL.md",[],[],{"skillNo":73,"name":74,"nameZh":6,"description":75,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":76,"categories":77,"keywords":78,"publishedAt":18},"01M78FHSBQ30GH1EMSBA8NJZNX","verifiability-first","Turn the task into something verifiable and declare the success criterion before seeing results. Use when the request is vague about \"done\" (\"make it work\", \"improve this\"), when writing tests, when fixing a bug (write the failing reproduction first), or when comparing two approaches. Critical before any measurement or benchmark, where deciding the bar after seeing the numbers guarantees a \"success\".","https:\u002F\u002Fgithub.com\u002Fsongjiun10-collab\u002FHncs\u002Fblob\u002Fmain\u002Fsenior-engineer-mindset\u002Fverifiability-first\u002FSKILL.md",[],[],{"skillNo":80,"name":81,"nameZh":6,"description":82,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":83,"categories":84,"keywords":85,"publishedAt":18},"01M78FHN4P77KF3T47M1NBT7XA","threat-and-scale-check","Check trust boundaries, scale behavior, and layered safety. Use when handling user input, external API responses, files, or auth; when designing permissions, validation, or any safety mechanism; when data volume could grow; or when writing loops that touch I\u002FO. Also use when reviewing whether a single safeguard is the only thing standing between a mistake and a disaster.","https:\u002F\u002Fgithub.com\u002Fsongjiun10-collab\u002FHncs\u002Fblob\u002Fmain\u002Fsenior-engineer-mindset\u002Fthreat-and-scale-check\u002FSKILL.md",[],[],{"skillNo":87,"name":88,"nameZh":6,"description":89,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":90,"categories":91,"keywords":92,"publishedAt":18},"01M78FHGGZKPTBRVWJK83X18CY","surgical-change","Keep every changed line traceable to the request. Use when editing existing code, especially code you didn't write; when tempted to clean up something you noticed while reading; when a diff is growing beyond the ask; or when deciding whether to delete code that looks dead.","https:\u002F\u002Fgithub.com\u002Fsongjiun10-collab\u002FHncs\u002Fblob\u002Fmain\u002Fsenior-engineer-mindset\u002Fsurgical-change\u002FSKILL.md",[],[]]