{"_id":"@codecell-germany/gemini-ocr-agent-skill","_rev":"3-9249b92ac4a0aa2092cdabe691f91ee7","name":"@codecell-germany/gemini-ocr-agent-skill","dist-tags":{"latest":"0.2.1"},"versions":{"0.1.0":{"name":"@codecell-germany/gemini-ocr-agent-skill","version":"0.1.0","keywords":["ocr","gemini","gemini-api","cli","skill","agents","markdown","pdf","image-ocr","codex"],"author":{"name":"Nikolas Gottschol"},"license":"MIT","_id":"@codecell-germany/gemini-ocr-agent-skill@0.1.0","maintainers":[{"name":"codecell-germany","email":"admin@codecell.de"}],"homepage":"https://github.com/codecell-germany/gemini-pdf-img-ocr-agent-skill#readme","bugs":{"url":"https://github.com/codecell-germany/gemini-pdf-img-ocr-agent-skill/issues"},"bin":{"gemini-ocr":"dist/index.js","gemini-ocr-skill":"dist/installer.js"},"dist":{"shasum":"48d97788aa073490b6eb883e3fb36cecf2bda376","tarball":"https://registry.npmjs.org/@codecell-germany/gemini-ocr-agent-skill/-/gemini-ocr-agent-skill-0.1.0.tgz","fileCount":61,"integrity":"sha512-XX9jRYPO2NyEK3FYQLffYxCFcUuZsWJxtuRIpy1ZM3p+Q+TJfDs4c7uKIbLfpwLDzM5TId4LSVT2h5q/lzIcjA==","signatures":[{"sig":"MEQCIHcar2KFWXAfgL3ahYTCG4mMvgmlwC5iCCKRJ32fYfx9AiBt3Zb+UVGNPGxMuiBVuipqa7e4jVPr4qbheO81ffZE9w==","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":163457},"engines":{"node":">=20"},"gitHead":"3f16c0fff0c79d39af0cc5b0fd30c2e322053533","private":false,"scripts":{"test":"vitest run tests/unit","build":"npm run clean && tsc -p tsconfig.json && node scripts/ensure-executable.js","clean":"node scripts/clean.js","prepack":"npm run build && npm run test && node scripts/check-package-contents.js","test:unit":"vitest run tests/unit","typecheck":"tsc --noEmit"},"_npmUser":{"name":"codecell-germany","email":"admin@codecell.de"},"repository":{"url":"git+https://github.com/codecell-germany/gemini-pdf-img-ocr-agent-skill.git","type":"git"},"_npmVersion":"11.12.1","description":"Agent-friendly Gemini OCR CLI and Codex skill for reliable OCR on PDFs and images through one Gemini-powered workflow","directories":{},"_nodeVersion":"25.9.0","dependencies":{"commander":"14.0.3","pdfjs-dist":"5.6.205","@google/genai":"1.49.0"},"publishConfig":{"access":"public"},"_hasShrinkwrap":false,"devDependencies":{"vitest":"4.1.4","typescript":"6.0.2","@types/node":"25.6.0"},"_npmOperationalInternal":{"tmp":"tmp/gemini-ocr-agent-skill_0.1.0_1775922477820_0.005045550982556657","host":"s3://npm-registry-packages-npm-production"}},"0.2.0":{"name":"@codecell-germany/gemini-ocr-agent-skill","version":"0.2.0","keywords":["ocr","gemini","gemini-api","cli","skill","agents","markdown","pdf","image-ocr","codex"],"author":{"name":"Nikolas Gottschol"},"license":"MIT","_id":"@codecell-germany/gemini-ocr-agent-skill@0.2.0","maintainers":[{"name":"codecell-germany","email":"admin@codecell.de"}],"homepage":"https://github.com/codecell-germany/gemini-pdf-img-ocr-agent-skill#readme","bugs":{"url":"https://github.com/codecell-germany/gemini-pdf-img-ocr-agent-skill/issues"},"bin":{"gemini-ocr":"dist/index.js","gemini-ocr-skill":"dist/installer.js"},"dist":{"shasum":"d368735161241d88fc057b4f099c47b76d208694","tarball":"https://registry.npmjs.org/@codecell-germany/gemini-ocr-agent-skill/-/gemini-ocr-agent-skill-0.2.0.tgz","fileCount":61,"integrity":"sha512-ghO4fym/8hFEmw/FIEwfmQvoJRLRzt97k2Q5ub+Xq6aIJCoJ4f64xlcEAZGpIGaJe1FJ/0KzBNNtD+Mkkt7Szg==","signatures":[{"sig":"MEQCIHVAVC3625Xl0d5hP3Z2AiJxo047A+3PcTaEZhSjGBURAiA/bxmEYvubDzGWm2K5uftjwvMJMYpIZkf0Ku4L0DwpBQ==","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":183476},"engines":{"node":">=20"},"gitHead":"0cc373c14d8e0514e89bce10bc20ba5f53bb3e3e","private":false,"scripts":{"test":"vitest run tests/unit","build":"npm run clean && tsc -p tsconfig.json && node scripts/ensure-executable.js","clean":"node scripts/clean.js","prepack":"npm run build && npm run test && node scripts/check-package-contents.js","test:unit":"vitest run tests/unit","typecheck":"tsc --noEmit"},"_npmUser":{"name":"codecell-germany","email":"admin@codecell.de"},"repository":{"url":"git+https://github.com/codecell-germany/gemini-pdf-img-ocr-agent-skill.git","type":"git"},"_npmVersion":"11.12.1","description":"Agent-friendly Gemini file analysis CLI and Codex skill for promptable image/PDF workflows with a document-focused OCR mode","directories":{},"_nodeVersion":"25.9.0","dependencies":{"commander":"14.0.3","pdfjs-dist":"5.6.205","@google/genai":"1.49.0"},"publishConfig":{"access":"public"},"_hasShrinkwrap":false,"devDependencies":{"vitest":"4.1.4","typescript":"6.0.2","@types/node":"25.6.0"},"_npmOperationalInternal":{"tmp":"tmp/gemini-ocr-agent-skill_0.2.0_1776330137196_0.058281932476270004","host":"s3://npm-registry-packages-npm-production"}},"0.2.1":{"name":"@codecell-germany/gemini-ocr-agent-skill","version":"0.2.1","private":false,"description":"Agent-friendly Gemini file analysis CLI and Codex skill for promptable image/PDF workflows with a document-focused OCR mode","license":"MIT","author":{"name":"Nikolas Gottschol"},"repository":{"type":"git","url":"git+https://github.com/codecell-germany/gemini-pdf-img-ocr-agent-skill.git"},"homepage":"https://github.com/codecell-germany/gemini-pdf-img-ocr-agent-skill#readme","bugs":{"url":"https://github.com/codecell-germany/gemini-pdf-img-ocr-agent-skill/issues"},"bin":{"gemini-ocr":"dist/index.js","gemini-ocr-skill":"dist/installer.js"},"scripts":{"clean":"node scripts/clean.js","build":"npm run clean && tsc -p tsconfig.json && node scripts/ensure-executable.js","typecheck":"tsc --noEmit","test":"vitest run tests/unit","test:unit":"vitest run tests/unit","prepack":"npm run build && npm run test && node scripts/check-package-contents.js"},"publishConfig":{"access":"public"},"engines":{"node":">=20"},"keywords":["ocr","gemini","gemini-api","cli","skill","agents","markdown","pdf","image-ocr","codex"],"dependencies":{"@google/genai":"1.49.0","commander":"14.0.3","pdfjs-dist":"5.6.205"},"devDependencies":{"@types/node":"25.6.0","typescript":"6.0.2","vitest":"4.1.4"},"gitHead":"9c06c4f48130c9420ad857f54f03560cbd321f9f","_id":"@codecell-germany/gemini-ocr-agent-skill@0.2.1","_nodeVersion":"25.9.0","_npmVersion":"11.12.1","dist":{"integrity":"sha512-sQUuFMeqpJACQTRLcDSsA0iaey0E/Pjk3+XACPNqZoEJFTXo/OlR9MdssY3x2w5WIzcgBE//wnlGuGwCM9RrDg==","shasum":"0ad4eca7aaffe4769259a788c8ef2a13f07e599b","tarball":"https://registry.npmjs.org/@codecell-germany/gemini-ocr-agent-skill/-/gemini-ocr-agent-skill-0.2.1.tgz","fileCount":61,"unpackedSize":192311,"signatures":[{"keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U","sig":"MEQCICxlUT8b3j4p6azS1UdNQ81KZE50pQndS9dO3UM3ssU6AiBj97ta7uoxtnTTqMu686Qzv9kS16vzp38AgV6Cv8P7tg=="}]},"_npmUser":{"name":"codecell-germany","email":"admin@codecell.de"},"directories":{},"maintainers":[{"name":"codecell-germany","email":"admin@codecell.de"}],"_npmOperationalInternal":{"host":"s3://npm-registry-packages-npm-production","tmp":"tmp/gemini-ocr-agent-skill_0.2.1_1776334473143_0.9441697289268507"},"_hasShrinkwrap":false}},"time":{"created":"2026-04-11T15:47:57.716Z","modified":"2026-04-16T10:14:33.392Z","0.1.0":"2026-04-11T15:47:57.974Z","0.2.0":"2026-04-16T09:02:17.389Z","0.2.1":"2026-04-16T10:14:33.285Z"},"bugs":{"url":"https://github.com/codecell-germany/gemini-pdf-img-ocr-agent-skill/issues"},"author":{"name":"Nikolas Gottschol"},"license":"MIT","homepage":"https://github.com/codecell-germany/gemini-pdf-img-ocr-agent-skill#readme","keywords":["ocr","gemini","gemini-api","cli","skill","agents","markdown","pdf","image-ocr","codex"],"repository":{"type":"git","url":"git+https://github.com/codecell-germany/gemini-pdf-img-ocr-agent-skill.git"},"description":"Agent-friendly Gemini file analysis CLI and Codex skill for promptable image/PDF workflows with a document-focused OCR mode","maintainers":[{"name":"codecell-germany","email":"admin@codecell.de"}],"readme":"# Gemini OCR Agent Skill\n\n---\n\n# English\n\n## Purpose\n\n`gemini-ocr-agent-skill` gives agents a dedicated Gemini file-analysis surface instead of forcing everything through the host model’s built-in vision behavior.\n\nIt is designed for two complementary workflows:\n\n- promptable image/PDF analysis with free-text output\n- document-focused OCR where PDFs and images can still be converted into faithful Markdown\n\nPublic release surfaces:\n\n- GitHub repo: `codecell-germany/gemini-pdf-img-ocr-agent-skill`\n- npm package: `@codecell-germany/gemini-ocr-agent-skill`\n- CLI binary: `gemini-ocr`\n- installer binary: `gemini-ocr-skill`\n- skill name: `gemini-ocr-cli`\n\n## What the package contains\n\n- A global CLI binary: `gemini-ocr`\n- A skill installer binary: `gemini-ocr-skill`\n- A Codex skill payload under `skills/gemini-ocr-cli/`\n- Onboarding and command references for first-run agent workflows\n- A generic `analyze-file` path for promptable file understanding\n- Preset document-mode commands for Markdown-oriented OCR\n\n## Installation\n\n### 1. Install the CLI globally\n\n```bash\nnpm install -g @codecell-germany/gemini-ocr-agent-skill\n```\n\n### 2. Verify the public binaries\n\n```bash\ngemini-ocr --help\ngemini-ocr-skill --help\n```\n\n### 3. Install the skill payload\n\n```bash\ngemini-ocr-skill install --force\n```\n\n## Canonical first-run sequence\n\n### 1. Verify the public CLI\n\n```bash\ngemini-ocr --help\n```\n\n### 2. Check the environment\n\n```bash\ngemini-ocr doctor --json\n```\n\n### 3. If setup is incomplete, print the setup guide\n\n```bash\ngemini-ocr setup --language en\n```\n\n### 4. Export one supported Gemini API key env var\n\n```bash\nexport GEMINI_API_KEY=\"...\"\n```\n\nSupported alias:\n\n```bash\nexport GOOGLE_GENERATIVE_AI_API_KEY=\"...\"\n```\n\nFallback alias:\n\n```bash\nexport GOOGLE_API_KEY=\"...\"\n```\n\n### 5. Re-run the environment check\n\n```bash\ngemini-ocr doctor --json\n```\n\n### 6. Run the real command\n\n```bash\ngemini-ocr analyze-file /absolute/path/to/image.png --prompt \"Describe this image precisely.\"\ngemini-ocr analyze-file /absolute/path/to/image.png --prompt \"Describe this image precisely.\" --reasoning-level low\ngemini-ocr scan-pdf /absolute/path/to/document.pdf\n```\n\n## Command overview\n\n- `gemini-ocr analyze-file <path> --prompt <text> [--output <path>] [--model <id>] [--reasoning-level minimal|low|medium|high]`\n- `gemini-ocr setup --language en|de`\n- `gemini-ocr doctor --json [--reasoning-level minimal|low|medium|high]`\n- `gemini-ocr scan-image <path> [--format markdown|json] [--output <path>] [--model <id>] [--reasoning-level minimal|low|medium|high]`\n- `gemini-ocr scan-pdf <path> [--pdf-mode auto|native|raster] [--dpi <n>] [--format markdown|json] [--reasoning-level minimal|low|medium|high]`\n- `gemini-ocr-skill install --force`\n- `gemini-ocr-skill uninstall`\n- `gemini-ocr-skill doctor`\n\n## Model policy\n\nAs of April 10, 2026, the current default policy is:\n\n- preferred default: `gemini-3.1-flash-lite-preview`\n- fallback: `gemini-3-flash-preview`\n- override via `--model` or `GEMINI_OCR_MODEL`\n- reasoning override via `--reasoning-level` or `GEMINI_OCR_REASONING_LEVEL`\n- if unset, Gemini 3 uses its model-default thinking behavior\n\n`doctor` prints the resolved policy so agents can see what will be used before running OCR.\n\n## Product modes\n\n- `analyze-file`\n  - accepts a local image or PDF\n  - requires `--prompt`\n  - returns free text\n  - uses the generic Gemini file-analysis path\n  - supports explicit Gemini thinking control via `--reasoning-level`\n- `scan-image`\n  - keeps the document/OCR-oriented image preset\n  - defaults to Markdown output\n- `scan-pdf`\n  - keeps the document/OCR-oriented PDF preset\n  - defaults to Markdown output\n  - retains the PDF-native vs raster logic\n\n## Document-mode behavior\n\n- Default output format is Markdown.\n- `--format json` returns the full typed OCR object for agent workflows.\n- PDF mode defaults to `auto`.\n- Default timeout is `180000` ms.\n- Native PDF requests get a higher adaptive timeout for large or many-page PDFs unless you set `--timeout-ms` explicitly.\n- `auto` samples the PDF text layer and prefers native PDF upload for born-digital and mixed PDFs.\n- When a local PDF text layer exists, the CLI also attaches it to the Gemini request so wording stays closer to the original while Gemini focuses on structure and tables.\n- If `pdftoppm` is available and the sampled PDF looks scanned, `auto` can rasterize pages first.\n- If a native `auto` PDF request aborts or times out, the CLI retries with raster page OCR when `pdftoppm` is available.\n- `--pdf-mode raster` requires a working `pdftoppm` binary.\n- `analyze-file` does not use the document JSON schema; it returns direct text from Gemini.\n\n## Agent guardrails\n\n- Use the public CLI names `gemini-ocr` and `gemini-ocr-skill`.\n- Do not bypass the product surface with `node dist/index.js`.\n- Do not call hidden runtime paths such as `~/.codex/tools/gemini-ocr-cli/dist/index.js`.\n- Run `doctor --json` in every new shell before the first real OCR task.\n- Use `analyze-file` for general promptable file understanding.\n- Use `scan-pdf` or `scan-image` when you explicitly want the document/OCR preset.\n- Use `--reasoning-level minimal|low|medium|high` when you need explicit latency vs. reasoning-depth control.\n\n## Privacy and remote processing\n\nThis tool sends file inputs to Gemini for remote processing.\n\nDo not use it for documents that must not leave the local machine unless that remote-processing policy is acceptable for the environment.\n\nThe CLI does not persist API keys.\nSecrets should stay in shell environment variables such as `GEMINI_API_KEY`, `GOOGLE_GENERATIVE_AI_API_KEY`, or `GOOGLE_API_KEY`.\n\n## Supported file inputs\n\nThe CLI currently supports:\n\n- PDFs\n- PNG\n- JPEG / JPG\n- WEBP\n- GIF\n- BMP\n- TIFF / TIF\n- HEIC\n- HEIF\n\n## Skills ecosystem\n\nRepository listing check:\n\n```bash\nnpx -y skills add codecell-germany/gemini-pdf-img-ocr-agent-skill -l\n```\n\nGlobal install example:\n\n```bash\nnpx -y skills add codecell-germany/gemini-pdf-img-ocr-agent-skill -g --skill gemini-ocr-cli -a '*' -y\n```\n\n## Release verification\n\n```bash\nnpm run build\nnpm test\nnpm whoami --registry=https://registry.npmjs.org/\nnpm pack\n```\n\nLocal tarball smoke test:\n\n```bash\nTMP=\"$(mktemp -d)\"\ncd \"$TMP\"\nnpx -y -p /absolute/path/to/package.tgz gemini-ocr --help\nnpx -y -p /absolute/path/to/package.tgz gemini-ocr-skill install --codex-home \"$TMP/codex\" --force\n\"$TMP/codex/bin/gemini-ocr\" --help\n```\n\n## References\n\n- Skill overview: `skills/gemini-ocr-cli/references/overview.md`\n- Agent onboarding: `skills/gemini-ocr-cli/references/agent-onboarding.md`\n- OCR first run: `skills/gemini-ocr-cli/references/ocr-first-run.md`\n- Command cheat sheet: `skills/gemini-ocr-cli/references/command-cheatsheet.md`\n- Architecture notes: `knowledge/ARCHITECTURE.md`\n- Known limitations: `knowledge/KNOWN_LIMITATIONS.md`\n- Release checklist: `knowledge/RELEASE_CHECKLIST.md`\n- Skill publishing guide: `knowledge/SKILL_PUBLISHING_GUIDE.md`\n\n---\n\n# Deutsch\n\n## Zweck\n\n`gemini-ocr-agent-skill` gibt Agenten eine dedizierte Gemini-Dateianalyse-Oberfläche, statt alles durch das eingebaute Vision-Verhalten des Host-Modells zu erzwingen.\n\nDas Paket ist für zwei zusammengehörige Arbeitsweisen gedacht:\n\n- promptbare Bild- und PDF-Analyse mit Freitext-Ausgabe\n- dokumentorientierte OCR, bei der Bilder und PDFs weiterhin möglichst originalgetreu in Markdown überführt werden können\n\nÖffentliche Release-Oberflächen:\n\n- GitHub-Repo: `codecell-germany/gemini-pdf-img-ocr-agent-skill`\n- npm-Paket: `@codecell-germany/gemini-ocr-agent-skill`\n- CLI-Binary: `gemini-ocr`\n- Installer-Binary: `gemini-ocr-skill`\n- Skill-Name: `gemini-ocr-cli`\n\n## Was das Paket enthält\n\n- Ein globales CLI-Binary: `gemini-ocr`\n- Ein Installer-Binary für den Skill: `gemini-ocr-skill`\n- Einen Codex-Skill unter `skills/gemini-ocr-cli/`\n- Onboarding- und Befehlsreferenzen für First-Run-Agenten-Workflows\n- Einen generischen `analyze-file`-Pfad für promptbare Dateiverarbeitung\n- Dokumentmodus-Kommandos für Markdown-orientierte OCR\n\n## Installation\n\n### 1. CLI global installieren\n\n```bash\nnpm install -g @codecell-germany/gemini-ocr-agent-skill\n```\n\n### 2. Öffentliche Binaries prüfen\n\n```bash\ngemini-ocr --help\ngemini-ocr-skill --help\n```\n\n### 3. Skill-Payload installieren\n\n```bash\ngemini-ocr-skill install --force\n```\n\n## Kanonischer First-Run-Ablauf\n\n```bash\ngemini-ocr --help\ngemini-ocr doctor --json\ngemini-ocr setup --language de\nexport GEMINI_API_KEY=\"...\"\n# oder:\nexport GOOGLE_GENERATIVE_AI_API_KEY=\"...\"\n# oder:\nexport GOOGLE_API_KEY=\"...\"\n# optional:\nexport GEMINI_OCR_REASONING_LEVEL=\"medium\"\ngemini-ocr doctor --json\ngemini-ocr analyze-file /absoluter/pfad/zur/datei.png --prompt \"Beschreibe das Bild präzise.\"\ngemini-ocr scan-pdf /absoluter/pfad/zum/dokument.pdf\n```\n\n## Agenten-Guardrails\n\n- Nutze die öffentlichen CLI-Namen `gemini-ocr` und `gemini-ocr-skill`.\n- Nutze keine repo-lokalen `dist/`-Entrypoints.\n- Nutze keine versteckten Runtime-Pfade wie `~/.codex/tools/gemini-ocr-cli/dist/index.js`.\n- Führe in jeder neuen Shell zuerst `gemini-ocr doctor --json` aus.\n- Nutze `analyze-file` für allgemeine promptbare Bild-/PDF-Analyse.\n- Nutze `scan-pdf` oder `scan-image`, wenn du bewusst den Dokument-/OCR-Preset willst.\n- `analyze-file` gibt standardmäßig Freitext zurück.\n- Im Dokumentmodus ist die Standardausgabe Markdown.\n- `--format json` gibt das vollständige strukturierte OCR-Objekt zurück.\n- PDF-Standardmodus ist `auto`.\n- Standard-Timeout ist `180000` ms.\n- Native PDF-Requests bekommen ohne explizites `--timeout-ms` ein höheres adaptives Timeout für große oder mehrseitige PDFs.\n- Bei born-digital PDFs hängt das CLI die lokal extrahierte Textschicht zusätzlich an den Gemini-Request, damit die Wortlaute stabiler bleiben und Gemini sich stärker auf Struktur und Tabellen konzentriert.\n- Wenn ein nativer `auto`-PDF-Request abbricht oder ins Timeout läuft, versucht das CLI automatisch Raster-OCR, sofern `pdftoppm` verfügbar ist.\n- `analyze-file` nutzt kein Dokument-JSON-Schema, sondern gibt direkten Text von Gemini zurück.\n- Das Reasoning-Level kann über `--reasoning-level minimal|low|medium|high` oder `GEMINI_OCR_REASONING_LEVEL` gesetzt werden.\n- API-Schlüssel werden nicht persistiert.\n- Dokumente werden zur Verarbeitung an Gemini übertragen.\n\n## Modellpolitik\n\n- Standardmodell: `gemini-3.1-flash-lite-preview`\n- Fallback: `gemini-3-flash-preview`\n- Override: `--model` oder `GEMINI_OCR_MODEL`\n- Optionales Reasoning-Level: `--reasoning-level` oder `GEMINI_OCR_REASONING_LEVEL`\n- Ohne Override nutzt Gemini 3 das modellseitige Standard-Thinking-Verhalten\n\n## Unterstützte Dateitypen\n\n- PDF\n- PNG\n- JPEG / JPG\n- WEBP\n- GIF\n- BMP\n- TIFF / TIF\n- HEIC\n- HEIF\n\n## skills.sh\n\n```bash\nnpx -y skills add codecell-germany/gemini-pdf-img-ocr-agent-skill -l\nnpx -y skills add codecell-germany/gemini-pdf-img-ocr-agent-skill -g --skill gemini-ocr-cli -a '*' -y\n```\n\n## Release-Verifikation\n\n```bash\nnpm run build\nnpm test\nnpm whoami --registry=https://registry.npmjs.org/\nnpm pack\n```\n\nLokaler Tarball-Smoke-Test:\n\n```bash\nTMP=\"$(mktemp -d)\"\ncd \"$TMP\"\nnpx -y -p /absoluter/pfad/zum/paket.tgz gemini-ocr --help\nnpx -y -p /absoluter/pfad/zum/paket.tgz gemini-ocr-skill install --codex-home \"$TMP/codex\" --force\n\"$TMP/codex/bin/gemini-ocr\" --help\n```\n\n## Referenzen\n\n- Übersicht: `skills/gemini-ocr-cli/references/overview.md`\n- Agenten-Onboarding: `skills/gemini-ocr-cli/references/agent-onboarding.md`\n- OCR-First-Run: `skills/gemini-ocr-cli/references/ocr-first-run.md`\n- Befehlsübersicht: `skills/gemini-ocr-cli/references/command-cheatsheet.md`\n- Architektur: `knowledge/ARCHITECTURE.md`\n- Bekannte Einschränkungen: `knowledge/KNOWN_LIMITATIONS.md`\n- Release-Checkliste: `knowledge/RELEASE_CHECKLIST.md`\n- Skill-Publishing-Guide: `knowledge/SKILL_PUBLISHING_GUIDE.md`\n","readmeFilename":"README.md"}