{"id":48379495,"url":"https://github.com/asaf-dahan/super-skill","last_synced_at":"2026-04-11T16:15:17.013Z","repository":{"id":348965596,"uuid":"1200474796","full_name":"Asaf-Dahan/super-skill","owner":"Asaf-Dahan","description":"A portable, AI-native intelligence layer that gives any language model full context, methodology, and operational capability over a defined domain.","archived":false,"fork":false,"pushed_at":"2026-04-03T15:33:40.000Z","size":41,"stargazers_count":1,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2026-04-03T18:36:10.217Z","etag":null,"topics":["agent-skills","ai-agent","claude-code","domain-experts","knowledge-management","llm","notebooklm-py","super-skill"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/Asaf-Dahan.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2026-04-03T13:06:26.000Z","updated_at":"2026-04-03T15:29:30.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/Asaf-Dahan/super-skill","commit_stats":null,"previous_names":["asaf-dahan/super-skill"],"tags_count":1,"template":true,"template_full_name":null,"purl":"pkg:github/Asaf-Dahan/super-skill","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Asaf-Dahan%2Fsuper-skill","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Asaf-Dahan%2Fsuper-skill/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Asaf-Dahan%2Fsuper-skill/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Asaf-Dahan%2Fsuper-skill/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/Asaf-Dahan","download_url":"https://codeload.github.com/Asaf-Dahan/super-skill/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Asaf-Dahan%2Fsuper-skill/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":31446531,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-04-05T15:22:31.103Z","status":"ssl_error","status_checked_at":"2026-04-05T15:22:00.205Z","response_time":75,"last_error":"SSL_connect returned=1 errno=0 peeraddr=140.82.121.5:443 state=error: unexpected eof while reading","robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":false,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["agent-skills","ai-agent","claude-code","domain-experts","knowledge-management","llm","notebooklm-py","super-skill"],"created_at":"2026-04-05T19:01:44.083Z","updated_at":"2026-04-05T19:01:45.087Z","avatar_url":"https://github.com/Asaf-Dahan.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Super Skill\n\nSuper Skill is an open framework for creating private domain experts\nfor you and your agents.\n\nEach Super Skill is an intelligence layer: a self-growing knowledge base\nthat your agents read before every task, and a living view of your domain\nexpert's brain through NotebookLM -- so you can understand, audit,\nand improve it over time.\n\nMultiple agents. One source of truth. Always current.\n\n---\n\n## What This Means in Practice\n\nYour agents stop starting from zero.\n\nEvery session, Claude Code reads your Super Skill before touching anything.\nIt knows what tools you run, what decisions you have made, what is waiting\nfor your approval, and what must never be re-litigated. It works from\nverified current state, not assumptions.\n\nYou stop explaining yourself in every conversation.\n\nNotebookLM holds the growing knowledge base for each domain.\nYou can listen to an audio overview of your own decisions,\ntake a quiz on your own architecture, and see a mind map of how\neverything connects. As you learn, you improve the Super Skill.\nAs the Super Skill improves, your agents become more accurate.\n\n---\n\n## Skill vs Super Skill\n\n| | Skill | Super Skill |\n|---|---|---|\n| What it is | A file | A living knowledge architecture |\n| Knows how to | Do a task | Understand a full domain |\n| Self-growing | No | Yes |\n| Detects drift | No | Yes |\n| Teaches the agent | Yes | Yes |\n| Teaches you | No | Yes -- audio, quiz, mind map |\n| Works with | Claude Code | Any agent, any model |\n| Time to activate | 1 minute | Under 15 minutes |\n\n---\n\n## When to Create a Super Skill\n\nAsk three questions before forking this template:\n\n1. Does this domain change externally without your control?\n   Tools release updates. APIs change. Pricing shifts.\n\n2. Are there decisions that must never be re-litigated?\n   Architectural choices, rejected alternatives, settled patterns.\n\n3. Does your agent need verified current state before every task?\n   Which plan, which version, which modules are active right now.\n\nTwo or more YES answers: create a Super Skill.\nFewer than two: a SKILL.md file is sufficient.\n\n---\n\n## How It Works\n\nPrompt 1 runs in three stages. Paste the entire block into Claude Code.\n\n```\nClone the template\n        |\nOpen Claude Code in the folder\n        |\nPaste Prompt 1 from ONBOARDING.md\n        |\nStage 1: Domain Layer Generation\n  Claude Code scans your environment\n  Claude Code asks at most 3 questions\n  Claude Code generates all 10 layer files\n        |\nStage 2: Expert Council Research\n  Claude Code proposes 10 real domain experts\n  You choose who joins your council\n        |\nStage 3: Expert Profile Generation\n  Claude Code generates expert profiles\n  Claude Code generates experts/COUNCIL.md with debates\n        |\nClaude Code creates or updates ~/.claude/CLAUDE.md\n(the global router that points to this Super Skill)\n        |\nEvery future session:\nClaude Code reads the router\nLoads SUMMARY.md from the right Super Skill\nStarts with full context in under 2 minutes\n```\n\n---\n\n## The Layer Architecture\n\nEvery Super Skill contains the same 10 files regardless of domain.\nThe structure is identical. The content is yours.\n\n```\nSUMMARY.md          session entry point -- 80 lines, always current\nSKILL.md            agent skills format entry point\nCLAUDE.md           claude code operating instructions\nAGENTS.md           instructions for any other agent\nCONTEXT.md          layer 0: identity, principles, scope\nDOMAIN_MAP.md       layer 1: domain structure and sub-domain relationships\nCURRENT_STATE.md    layer 2: verified current state\nEVALUATION.md       layer 3: framework for evaluating anything new\nDECISIONS.md        layer 4: decisions log and reasoning\nMONITORING.md       layer 5: drift detection sources\nLEARNING.md         layer 6: notebooklm notebook plan\nPENDING.md          layer 7: approval queue\nexperts/COUNCIL.md  layer X: expert council and debates\nexperts/[name].md   individual expert profiles\nnotebooks/          notebooklm-ready learning files\nscripts/            automation: feed, generate, check, sync, update_summary\nchangelog/          monthly detected changes\n```\n\n---\n\n## The Operating Principle\n\nThe model proposes.\nYou decide.\nThe Super Skill records.\nThe system executes.\n\nNo layer file changes without your approval.\nEvery proposed change goes to PENDING.md first.\nYou review, you decide, then and only then it is recorded.\n\n---\n\n## NotebookLM -- The Learning Layer\n\nNotebookLM is optional but transforms what a Super Skill can do.\n\nWithout NotebookLM: your agents read Markdown files from the repository.\nFull context, zero re-explanation, every session.\n\nWith NotebookLM: the knowledge base grows beyond static files.\nYou and your agents can add research, documents, and new sources.\nClaude proposes additions with your approval. You add directly in the notebook.\nThe notebook generates audio overviews, quizzes, and mind maps from your\nown domain knowledge. Token usage drops significantly because Claude queries\nthe notebook instead of loading full files.\n\nConnect it in Step 4 of ONBOARDING.md. Skip it and add it later if you prefer.\n\n---\n\n## Expert Council\n\nDuring activation, Claude Code researches your domain and proposes\n10 real, leading experts with verifiable published work. You choose\nwho joins your council. Each expert is assigned to the layer they\ncontribute most to and given a full profile with methodology,\nframeworks, red lines, and five questions they would ask you.\n\nWhen experts disagree, the disagreement is recorded as a debate in\nexperts/COUNCIL.md. Claude Code surfaces expert perspectives only\nwhen genuine friction exists -- not on routine tasks.\n\nIf an expert's position conflicts with a decision in DECISIONS.md,\nit goes to PENDING.md as an expert challenge for your review.\n\n---\n\n## Domain Types\n\nSuper Skill works for any domain where knowledge matters and decisions accumulate.\n\nTechnical: software infrastructure, database architecture, security frameworks,\nDevOps pipelines, product methodology, AI agent systems.\n\nProfessional: investment portfolio, legal review, medical documentation,\nmarketing strategy, real estate, financial modeling, operations.\n\nPersonal: home garden, fitness, nutrition, home renovation, language learning,\ntravel planning, personal finance.\n\nCraft: photography, woodworking, wine, cooking techniques, academic research.\n\nIf you can describe it, a Super Skill can master it.\n\n---\n\n## Get Started\n\nClone this repo. Open Claude Code. Paste Prompt 1 from ONBOARDING.md.\n\n  git clone https://github.com/Asaf-Dahan/super-skill my-super-skill-[domain]\n  cd my-super-skill-[domain]\n  claude .\n\nYour Super Skill will be live in under 15 minutes.\n\n---\n\n## Version\n\nSuper Skill v2.1.0\n\n## License\n\nMIT License -- (c) 2026 Gitit Inc - AI Architecture\n\nFork it. Build your own. Publish your Super Skills.\n\n---\n\n## Important Notes\n\nnotebooklm-py is an unofficial library that uses undocumented Google APIs.\nIt may change without notice. Suitable for personal and internal use.\nNot recommended for production systems serving external users.\n\nKeep domain-specific Super Skills in private repositories.\nNever commit API keys or tokens to any layer file.\nUse .env for all sensitive values.\n\nAll agent actions follow one rule:\nThe model proposes. You decide. Nothing changes without your approval.\n\n---\n\n## Created By\n\nSuper Skill was built by Asaf Dahan,\nAI Solutions Architect at Gitit Inc.\n\nWebsite: ai.asafid.com\nCompany: Gitit Inc\nGitHub: github.com/Asaf-Dahan\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fasaf-dahan%2Fsuper-skill","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fasaf-dahan%2Fsuper-skill","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fasaf-dahan%2Fsuper-skill/lists"}