[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7AAAYHB4W95TF4918PYG8Z2":3,"skill-same:01M7AAAYHB4W95TF4918PYG8Z2":59},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":24,"commitSha":25,"skillDirectory":26,"forksCount":27,"repositoryHtmlUrl":28,"repositoryHomepageUrl":6,"ownerHtmlUrl":29,"ownerBlogUrl":30,"ownerDescription":31,"updatedAt":32,"install":33,"resources":46},"01M7AAAYHB4W95TF4918PYG8Z2","generate-art",null,"Generate static artwork\u002Fillustrations for a repo with Nano Banana Pro (aka\n\"nano banana\", Gemini 3 Pro Image) via the bundled generate_art.py — craft a\ndetailed prompt, run the script, view the result, iterate. Use this whenever the\nuser wants to create, generate, or regenerate a visual asset for a project: an\narchitecture diagram, portrait, card\u002Fillustration, icon, splash, or any image\nthat lands as a PNG in the repo — even if they don't name the tool or just say\n\"make me an image \u002F a picture of X\". Do NOT trigger for application runtime image\nneeds, or for editing source code rather than producing an image file.","JasperHG90\u002Fshakedown","JasperHG90","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F8079803?v=4","01M7AAAKN62DYZ9XRCXAN0HGEF","Jasper Ginn","01M7AAAKP8D78B8ANVKEV7MRK8",1,"Apache-2.0","https:\u002F\u002Fgithub.com\u002FJasperHG90\u002Fshakedown\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fgenerate-art\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: generate-art\ndescription: |\n  Generate static artwork\u002Fillustrations for a repo with Nano Banana Pro (aka\n  \"nano banana\", Gemini 3 Pro Image) via the bundled generate_art.py — craft a\n  detailed prompt, run the script, view the result, iterate. Use this whenever the\n  user wants to create, generate, or regenerate a visual asset for a project: an\n  architecture diagram, portrait, card\u002Fillustration, icon, splash, or any image\n  that lands as a PNG in the repo — even if they don't name the tool or just say\n  \"make me an image \u002F a picture of X\". Do NOT trigger for application runtime image\n  needs, or for editing source code rather than producing an image file.\n---\n\n# Generate art\n\nBuild-time asset tooling. The engine is `scripts\u002Fgenerate_art.py`, bundled in\nthis skill; the skill body is the prompt-craft + iterate loop around it.\n\nThis skill is a **general** art generator — it knows *how* to generate, apply a\nstyle, and cast a character, but ships with no project-specific art. A project's\nhouse styles live in the *project* (`docs\u002Fstyles\u002F` by default, see \"Style\npresets\"), and the cast pool is just whatever portraits a project points\n`--cast-dir` at. Same skill, any repo.\n\n## Boundary — build-time only\n\nThis is build-time tooling, **not application runtime**. It produces static PNGs\nchecked into the repo; it should never run inside a shipped application or its\nrequest path. Keep the Gemini SDK and `GEMINI_API_KEY` out of runtime code paths.\nIf a project has rules constraining which LLMs may appear in runtime code, honour\nthem — this script's place is the build\u002Fasset pipeline, not the app.\n\n## Prerequisites\n\n- `GEMINI_API_KEY` (or `GOOGLE_API_KEY`) set in the environment. The script\n  reads it; do not hardcode keys.\n- `uv` available. The script is a self-contained PEP 723 script — `uv run`\n  installs `google-genai` + `pillow` into an ephemeral env. No project deps.\n\n## Run\n\nRun from the repo root (adjust the path to wherever the skill is installed):\n\n```\nuv run scripts\u002Fgenerate_art.py --prompt \"\u003Cprompt>\" --out \u003Cpath.png> [--banana pro] [--aspect 16:9] [--size 2K] [--ref existing.png]\n```\n\n- `--prompt` \u002F `--prompt-file` — one is required. For long prompts, write to a\n  temp file and use `--prompt-file` to avoid shell-quoting pain.\n- `--out` — required. Save to a **tracked** dir (be mindful of `.gitignore`).\n  See output conventions below.\n- `--banana` — `pro` (Gemini 3 Pro Image, default) or `flash` (Gemini 2.5 Flash\n  Image — cheaper\u002Ffaster, lower fidelity). Use `flash` for quick drafts\u002Ficons,\n  `pro` for hero\u002Fdiagram art and anything text-heavy or detail-critical.\n- `--model` — explicit model id; overrides `--banana` if a new variant ships.\n- `--aspect` — `1:1 3:2 2:3 3:4 4:3 4:5 5:4 9:16 16:9 21:9` (default `16:9`).\n- `--size` — `1K | 2K | 4K` (default `2K`). Pro only; ignored for `flash`.\n- `--style` — named style preset from `--styles-dir`. Appends the distilled\n  house-style block to the prompt. Prefer this over `--ref` for style matching —\n  see \"Style presets\" below.\n- `--styles-dir` — where project style presets live (default `docs\u002Fstyles`).\n- `--style-file` — ad-hoc style preset file; overrides `--style`.\n- `--cast` — opt-in: feature a real person's likeness in the art. `random`\n  picks one; or pass a name (file stem in `--cast-dir`). Off by default. Great\n  for people\u002Fprocess art where it's nice that the moment happens to a real\n  person. Combine with a `--style`.\n- `--cast-dir` — portrait dir for `--cast` (default `assets\u002Fportraits`).\n- `--ref` — reference image(s) for style\u002Fedit continuity; repeatable. Use to\n  edit an existing image, or to anchor on a specific asset when text alone isn't\n  enough. Note: passing a *character* portrait as `--ref` reproduces that person\n  — for generic art prefer `--style`; for *intentional* person reuse prefer\n  `--cast`.\n\n## Style presets\n\nStyle presets are **project content, not part of this skill.** The skill is a\ngeneral art generator; each project keeps its own house styles and points\n`--styles-dir` at them (default `docs\u002Fstyles\u002F`). A preset is a markdown file of\ndistilled style guidance that `--style \u003Cname>` appends to the prompt, so a house\nlook stays consistent across calls — without passing a character image as `--ref`\n(which tends to clone that specific person).\n\n**Use a preset:** add `--style \u003Cname>` to any call (resolves to\n`docs\u002Fstyles\u002F\u003Cname>.md`). The per-asset prompt still carries the *scene* and\ncomposition negatives (no text, no frame, aspect); the preset carries only the\n*look*.\n\n**Distil a new style** (the reusable \"capture the look once\" step):\n1. Look at 2-3 real reference images with the Read tool — the actual rendered art\n   you want to match, not a written spec (specs drift from reality).\n2. Write what you observe: medium\u002Frendering, linework, shading & light, palette,\n   background treatment, mood. Describe the *look*, not any scene or subject.\n   Keep composition\u002Fnegatives out — those are per-asset.\n3. Save it to the project's styles dir as `docs\u002Fstyles\u002F\u003Cname>.md` with an\n   `# Style: \u003Cname>` heading. (It lives in the project, not the skill.)\n4. Validate: generate one asset with `--style \u003Cname>` and no character `--ref`,\n   then Read the output and check it matches the references.\n\nKeep presets look-only and subject-agnostic so they compose with any prompt.\n\n**Generic vs. casting a real person.** Two deliberate paths, both opt-in:\n- **Generic in-style** (default\u002Fstandard) — `--style \u003Cname>`, no character ref.\n  The art matches the house look but depicts nobody in particular.\n- **Cast a real person** (option) — add `--cast random` (or `--cast \u003Cname>`) to\n  feature an actual portrait from `--cast-dir`. Nice for people\u002Fprocess art where\n  it lands that the moment happens to a real person. The script logs which\n  portrait it cast. This is an option, not the default — only reach for it when\n  reuse is wanted.\n\n## Output conventions\n\nSave assets to a tracked directory and pick the aspect that fits the asset. A\ncommon layout:\n\n- **Portraits** → `assets\u002Fportraits\u002F\u003Cname>.png`, `--aspect 3:4`. Match an\n  existing roster's style (pass an existing portrait via `--ref` or a `--style`).\n- **Diagrams \u002F architecture visuals** → `docs\u002Fassets\u002F`, `--aspect 16:9`.\n- **Card \u002F icon \u002F illustration art** → ask where it should live; default\n  `docs\u002Fassets\u002F`.\n\nConfirm the destination dir is not `.gitignore`d before writing.\n\n## Workflow\n\n1. **Clarify the asset** — what it depicts, where it lands, aspect, and whether\n   it must match an existing style (if so, use `--style`; distil a new preset\n   first if none fits).\n2. **Craft the prompt yourself.** Ground it in the project's design docs and any\n   relevant rules. Be specific about subject, layout, labels, palette, and style;\n   end with negatives (no logos, no paragraphs of text). Long prompts → temp\n   file + `--prompt-file`.\n3. **Run** the script.\n4. **View the result** with the Read tool on the output PNG and judge it\n   against what was asked — do not claim success without looking.\n5. **Iterate** on the prompt (or add `--ref`) until it's right. Show the user\n   the image and the final path.\n\n## Notes\n\n- Output is non-deterministic; same prompt varies run to run.\n- If the response has no image (safety filter \u002F refusal), the script exits with\n  an error — adjust the prompt and retry.\n","# Generate art\n\nBuild-time asset tooling. The engine is `scripts\u002Fgenerate_art.py`, bundled in\nthis skill; the skill body is the prompt-craft + iterate loop around it.\n\nThis skill is a **general** art generator — it knows *how* to generate, apply a\nstyle, and cast a character, but ships with no project-specific art. A project's\nhouse styles live in the *project* (`docs\u002Fstyles\u002F` by default, see \"Style\npresets\"), and the cast pool is just whatever portraits a project points\n`--cast-dir` at. Same skill, any repo.\n\n## Boundary — build-time only\n\nThis is build-time tooling, **not application runtime**. It produces static PNGs\nchecked into the repo; it should never run inside a shipped application or its\nrequest path. Keep the Gemini SDK and `GEMINI_API_KEY` out of runtime code paths.\nIf a project has rules constraining which LLMs may appear in runtime code, honour\nthem — this script's place is the build\u002Fasset pipeline, not the app.\n\n## Prerequisites\n\n- `GEMINI_API_KEY` (or `GOOGLE_API_KEY`) set in the environment. The script\n  reads it; do not hardcode keys.\n- `uv` available. The script is a self-contained PEP 723 script — `uv run`\n  installs `google-genai` + `pillow` into an ephemeral env. No project deps.\n\n## Run\n\nRun from the repo root (adjust the path to wherever the skill is installed):\n\n```\nuv run scripts\u002Fgenerate_art.py --prompt \"\u003Cprompt>\" --out \u003Cpath.png> [--banana pro] [--aspect 16:9] [--size 2K] [--ref existing.png]\n```\n\n- `--prompt` \u002F `--prompt-file` — one is required. For long prompts, write to a\n  temp file and use `--prompt-file` to avoid shell-quoting pain.\n- `--out` — required. Save to a **tracked** dir (be mindful of `.gitignore`).\n  See output conventions below.\n- `--banana` — `pro` (Gemini 3 Pro Image, default) or `flash` (Gemini 2.5 Flash\n  Image — cheaper\u002Ffaster, lower fidelity). Use `flash` for quick drafts\u002Ficons,\n  `pro` for hero\u002Fdiagram art and anything text-heavy or detail-critical.\n- `--model` — explicit model id; overrides `--banana` if a new variant ships.\n- `--aspect` — `1:1 3:2 2:3 3:4 4:3 4:5 5:4 9:16 16:9 21:9` (default `16:9`).\n- `--size` — `1K | 2K | 4K` (default `2K`). Pro only; ignored for `flash`.\n- `--style` — named style preset from `--styles-dir`. Appends the distilled\n  house-style block to the prompt. Prefer this over `--ref` for style matching —\n  see \"Style presets\" below.\n- `--styles-dir` — where project style presets live (default `docs\u002Fstyles`).\n- `--style-file` — ad-hoc style preset file; overrides `--style`.\n- `--cast` — opt-in: feature a real person's likeness in the art. `random`\n  picks one; or pass a name (file stem in `--cast-dir`). Off by default. Great\n  for people\u002Fprocess art where it's nice that the moment happens to a real\n  person. Combine with a `--style`.\n- `--cast-dir` — portrait dir for `--cast` (default `assets\u002Fportraits`).\n- `--ref` — reference image(s) for style\u002Fedit continuity; repeatable. Use to\n  edit an existing image, or to anchor on a specific asset when text alone isn't\n  enough. Note: passing a *character* portrait as `--ref` reproduces that person\n  — for generic art prefer `--style`; for *intentional* person reuse prefer\n  `--cast`.\n\n## Style presets\n\nStyle presets are **project content, not part of this skill.** The skill is a\ngeneral art generator; each project keeps its own house styles and points\n`--styles-dir` at them (default `docs\u002Fstyles\u002F`). A preset is a markdown file of\ndistilled style guidance that `--style \u003Cname>` appends to the prompt, so a house\nlook stays consistent across calls — without passing a character image as `--ref`\n(which tends to clone that specific person).\n\n**Use a preset:** add `--style \u003Cname>` to any call (resolves to\n`docs\u002Fstyles\u002F\u003Cname>.md`). The per-asset prompt still carries the *scene* and\ncomposition negatives (no text, no frame, aspect); the preset carries only the\n*look*.\n\n**Distil a new style** (the reusable \"capture the look once\" step):\n1. Look at 2-3 real reference images with the Read tool — the actual rendered art\n   you want to match, not a written spec (specs drift from reality).\n2. Write what you observe: medium\u002Frendering, linework, shading & light, palette,\n   background treatment, mood. Describe the *look*, not any scene or subject.\n   Keep composition\u002Fnegatives out — those are per-asset.\n3. Save it to the project's styles dir as `docs\u002Fstyles\u002F\u003Cname>.md` with an\n   `# Style: \u003Cname>` heading. (It lives in the project, not the skill.)\n4. Validate: generate one asset with `--style \u003Cname>` and no character `--ref`,\n   then Read the output and check it matches the references.\n\nKeep presets look-only and subject-agnostic so they compose with any prompt.\n\n**Generic vs. casting a real person.** Two deliberate paths, both opt-in:\n- **Generic in-style** (default\u002Fstandard) — `--style \u003Cname>`, no character ref.\n  The art matches the house look but depicts nobody in particular.\n- **Cast a real person** (option) — add `--cast random` (or `--cast \u003Cname>`) to\n  feature an actual portrait from `--cast-dir`. Nice for people\u002Fprocess art where\n  it lands that the moment happens to a real person. The script logs which\n  portrait it cast. This is an option, not the default — only reach for it when\n  reuse is wanted.\n\n## Output conventions\n\nSave assets to a tracked directory and pick the aspect that fits the asset. A\ncommon layout:\n\n- **Portraits** → `assets\u002Fportraits\u002F\u003Cname>.png`, `--aspect 3:4`. Match an\n  existing roster's style (pass an existing portrait via `--ref` or a `--style`).\n- **Diagrams \u002F architecture visuals** → `docs\u002Fassets\u002F`, `--aspect 16:9`.\n- **Card \u002F icon \u002F illustration art** → ask where it should live; default\n  `docs\u002Fassets\u002F`.\n\nConfirm the destination dir is not `.gitignore`d before writing.\n\n## Workflow\n\n1. **Clarify the asset** — what it depicts, where it lands, aspect, and whether\n   it must match an existing style (if so, use `--style`; distil a new preset\n   first if none fits).\n2. **Craft the prompt yourself.** Ground it in the project's design docs and any\n   relevant rules. Be specific about subject, layout, labels, palette, and style;\n   end with negatives (no logos, no paragraphs of text). Long prompts → temp\n   file + `--prompt-file`.\n3. **Run** the script.\n4. **View the result** with the Read tool on the output PNG and judge it\n   against what was asked — do not claim success without looking.\n5. **Iterate** on the prompt (or add `--ref`) until it's right. Show the user\n   the image and the final path.\n\n## Notes\n\n- Output is non-deterministic; same prompt varies run to run.\n- If the response has no image (safety filter \u002F refusal), the script exits with\n  an error — adjust the prompt and retry.\n",{"name":5,"description":7},"main","3ac632ac167bd480d8b07868b7001fc52e2a402c",".claude\u002Fskills\u002Fgenerate-art",0,"https:\u002F\u002Fgithub.com\u002FJasperHG90\u002Fshakedown","https:\u002F\u002Fgithub.com\u002FJasperHG90","https:\u002F\u002Fjasperhg90.github.io\u002F","Data Engineer @ Xebia Data. I am interested in data analysis, history, politics, education, policy making and writing. ","2026-09-21 04:18:57",{"source":8,"skillName":5,"commands":34,"prompt":44,"zipAvailable":45},[35,38,41],{"tool":36,"command":37},"npx","npx skills add JasperHG90\u002Fshakedown --skill generate-art",{"tool":39,"command":40},"bunx","bunx skills add JasperHG90\u002Fshakedown --skill generate-art",{"tool":42,"command":43},"pnpm","pnpm dlx skills add JasperHG90\u002Fshakedown --skill generate-art","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FJasperHG90\u002Fshakedown\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fgenerate-art\u002FSKILL.md",true,[47,53],{"path":48,"resourceType":49,"resourceTypeDesc":50,"sizeBytes":51,"sourceHtmlUrl":52},"README.md","REFERENCE","参考文档",2204,"https:\u002F\u002Fgithub.com\u002FJasperHG90\u002Fshakedown\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fgenerate-art\u002FREADME.md",{"path":54,"resourceType":55,"resourceTypeDesc":56,"sizeBytes":57,"sourceHtmlUrl":58},"scripts\u002Fgenerate_art.py","SCRIPT","脚本",8058,"https:\u002F\u002Fgithub.com\u002FJasperHG90\u002Fshakedown\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fgenerate-art\u002Fscripts\u002Fgenerate_art.py",[60,67,74,81,88,95],{"skillNo":61,"name":62,"nameZh":6,"description":63,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":64,"categories":65,"keywords":66,"publishedAt":19},"01M7AAFD8Q4CRZ5V34GB5655FX","fake-skill","A skill used only by shakedown's own offline tests.","https:\u002F\u002Fgithub.com\u002FJasperHG90\u002Fshakedown\u002Fblob\u002Fmain\u002Ftests\u002Ffake\u002Fskill\u002FSKILL.md",[],[],{"skillNo":68,"name":69,"nameZh":6,"description":70,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":71,"categories":72,"keywords":73,"publishedAt":19},"01M7AAFB03PDQMX0DY841ADDR9","shakedown-canary","Run the shakedown canary check. Use whenever someone asks to run the canary check, verify the harness, or confirm shakedown is wired up correctly.","https:\u002F\u002Fgithub.com\u002FJasperHG90\u002Fshakedown\u002Fblob\u002Fmain\u002Fsrc\u002Fshakedown\u002Fcanary\u002FSKILL.md",[],[],{"skillNo":75,"name":76,"nameZh":6,"description":77,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":78,"categories":79,"keywords":80,"publishedAt":19},"01M7AAF55DP6C2XSF7N23STFP3","create-cases","Author a shakedown cases file (`shakedowns\u002F\u003Cslug>.cases.toml`) for an existing agent skill: read the skill, turn what it promises into cases, supply fixtures for the commands it must not really run, and check the result with `shakedown case validate`. Use this whenever someone wants to measure, smoke-test, or add conformance cases for a skill, asks \"how do I test this skill\", points at a SKILL.md and wants cases written, or is editing a `.cases.toml`. Use it even when they only name the skill (\"write cases for connect-agent-to-hub\") without mentioning shakedown or TOML.","https:\u002F\u002Fgithub.com\u002FJasperHG90\u002Fshakedown\u002Fblob\u002Fmain\u002Fskills\u002Fcreate-cases\u002FSKILL.md",[],[],{"skillNo":82,"name":83,"nameZh":6,"description":84,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":85,"categories":86,"keywords":87,"publishedAt":19},"01M7AAEXCDG258QRVSSFQV0MV9","analyze-results","Read a shakedown report (`shakedown-report.json`) and say which file to open: the measured skill's own `SKILL.md`, the `shakedowns\u002F\u003Cslug>.cases.toml` holding its cases, the harness block in `shakedown.toml`, or a flag on the run itself. Use this whenever someone has run `shakedown case run` and wants to know what a number means or what to do about it, pastes a report or the PR comment rendered from one, asks why `skill_fired`, `tool_used`, `artifact_created` or `inputs_resolved` came out low, asks whether a rate is worth acting on, or says a conformance run failed and they cannot tell whose fault it is. Use it even when they only quote the percentage (\"my skill scores 60% on inputs_resolved, now what\") without naming shakedown or the report file.","https:\u002F\u002Fgithub.com\u002FJasperHG90\u002Fshakedown\u002Fblob\u002Fmain\u002Fskills\u002Fanalyze-results\u002FSKILL.md",[],[],{"skillNo":89,"name":90,"nameZh":6,"description":91,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":92,"categories":93,"keywords":94,"publishedAt":19},"01M7AAEMM68JP8A128WHMTJJTX","add-harness","Add a new agent harness (opencode, Gemini CLI, Codex, Amp, a local Ollama-backed runner, or any CLI that loads skills) to a shakedown config so its skill conformance can be measured, and diagnose a harness that `shakedown doctor` reports as not qualifying. Use this whenever someone wants shakedown to run against a harness it does not yet support, asks why a harness fails doctor, asks how to point an existing harness at a different model or provider, or is editing the `[harness.*]` block of a shakedown.toml. Use it even when they only name the tool (\"can this test opencode?\") without mentioning config or TOML.","https:\u002F\u002Fgithub.com\u002FJasperHG90\u002Fshakedown\u002Fblob\u002Fmain\u002Fskills\u002Fadd-harness\u002FSKILL.md",[],[],{"skillNo":96,"name":76,"nameZh":6,"description":77,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":97,"categories":98,"keywords":99,"publishedAt":19},"01M7AAEEKJAS0KAJRM74K7FEFJ","https:\u002F\u002Fgithub.com\u002FJasperHG90\u002Fshakedown\u002Fblob\u002Fmain\u002Fplugins\u002Fopencode\u002Fskills\u002Fcreate-cases\u002FSKILL.md",[],[]]