[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"extension-skill-0xmariowu-pubmed-zh-CN":3,"guides-for-0xmariowu-pubmed":696,"similar-k176j2rf6qje3y31sz6f1dxgk186m06v-zh-CN":697},{"_creationTime":4,"_id":5,"children":6,"community":7,"display":9,"evaluation":15,"identity":256,"isFallback":239,"parentExtension":261,"providers":262,"relations":267,"repo":269,"tags":693,"workflow":694},1778668068134.9573,"k176j2rf6qje3y31sz6f1dxgk186m06v",[],{"reviewCount":8},0,{"description":10,"installMethods":11,"name":13,"sourceUrl":14},"Use for biomedical and life-science literature searches when query expects PubMed articles, clinical trials, or medical research papers.",{"claudeCode":12},"0xmariowu/Autosearch","PubMed Search","https://github.com/0xmariowu/Autosearch",{"_creationTime":16,"_id":17,"extensionId":5,"locale":18,"result":19,"trustSignals":237,"workflow":254},1778668214554.111,"kn771sc1gey1wdxsjbjh7d85k586n9eg","en",{"checks":20,"evaluatedAt":196,"extensionSummary":197,"features":198,"nonGoals":203,"practices":209,"prerequisites":212,"promptVersionExtension":213,"promptVersionScoring":214,"purpose":215,"rationale":216,"score":217,"summary":218,"tags":219,"targetMarket":225,"tier":226,"useCases":227,"workflow":232},[21,26,29,32,36,39,43,48,51,54,58,62,65,69,72,75,78,81,84,87,91,95,100,104,108,111,114,117,121,124,127,130,133,136,139,143,147,151,154,158,161,164,167,170,174,177,180,183,186,189,193],{"category":22,"check":23,"severity":24,"summary":25},"Practical Utility","Problem relevance","pass","The displayed description clearly states the problem: searching biomedical and life-science literature for specific types of articles like PubMed articles, clinical trials, or medical research papers.",{"category":22,"check":27,"severity":24,"summary":28},"Unique selling proposition","The skill provides direct access to specialized biomedical literature databases (PubMed), offering value beyond generic web search or LLM training data, enabling targeted research.",{"category":22,"check":30,"severity":24,"summary":31},"Production readiness","The skill is production-ready, providing access to PubMed which is a well-established resource, and the installation and usage are clearly documented. The MIT license and active maintenance suggest readiness.",{"category":33,"check":34,"severity":24,"summary":35},"Scope","Single responsibility principle","The skill focuses exclusively on biomedical and life-science literature searches via PubMed, adhering to a single responsibility principle.",{"category":33,"check":37,"severity":24,"summary":38},"Description quality","The displayed description accurately reflects the skill's purpose of searching for biomedical and life-science literature, specifically mentioning PubMed articles, clinical trials, and medical research papers.",{"category":40,"check":41,"severity":24,"summary":42},"Invocation","Scoped tools","The skill uses a single method 'api_search' which is scoped to searching PubMed, aligning with the verb-noun specialization principle.",{"category":44,"check":45,"severity":46,"summary":47},"Documentation","Configuration & parameter reference","not_applicable","The skill does not appear to have any configurable parameters or options beyond its core search functionality, making this check not applicable.",{"category":33,"check":49,"severity":24,"summary":50},"Tool naming","The single exposed method 'api_search' is descriptive and relevant to the tool's function.",{"category":33,"check":52,"severity":24,"summary":53},"Minimal I/O surface","The 'api_search' method likely takes a query as input and returns structured search results, implying a minimal I/O surface tailored to its function.",{"category":55,"check":56,"severity":24,"summary":57},"License","License usability","The project is licensed under the MIT license, which is a permissive open-source license, and this is clearly indicated in the LICENSE file and README.",{"category":59,"check":60,"severity":24,"summary":61},"Maintenance","Commit recency","The last commit was on May 11, 2026, which is recent, indicating active maintenance.",{"category":59,"check":63,"severity":24,"summary":64},"Dependency Management","The project has a lockfile (implied by hasLockfile: true) and no complex third-party dependencies are evident in the provided files, suggesting good dependency management.",{"category":66,"check":67,"severity":24,"summary":68},"Security","Secret Management","The skill does not appear to use any secrets as it accesses PubMed directly via its E-utilities API without requiring an API key.",{"category":66,"check":70,"severity":24,"summary":71},"Injection","The skill is designed to search literature and does not appear to load or execute external code or data that could be untrusted. It uses a well-defined API.",{"category":66,"check":73,"severity":24,"summary":74},"Transitive Supply-Chain Grenades","The skill relies on a public API and does not fetch external code or data at runtime, mitigating supply-chain risks.",{"category":66,"check":76,"severity":24,"summary":77},"Sandbox Isolation","The skill operates by making API calls to PubMed and does not modify files outside its own project folder, maintaining sandbox isolation.",{"category":66,"check":79,"severity":24,"summary":80},"Sandbox escape primitives","No detached process spawns or deny-retry loops were observed in the provided code snippets.",{"category":66,"check":82,"severity":24,"summary":83},"Data Exfiltration","The skill only queries a public API for literature search results and does not submit confidential data to any third party.",{"category":66,"check":85,"severity":24,"summary":86},"Hidden Text Tricks","The bundled content appears to be free of hidden-steering tricks, and descriptions are clean printable ASCII.",{"category":88,"check":89,"severity":24,"summary":90},"Hooks","Opaque code execution","The skill's implementation ('methods/api_search.py') is expected to be plain, readable source code and not involve obfuscation techniques like base64 payloads or eval.",{"category":92,"check":93,"severity":24,"summary":94},"Portability","Structural Assumption","The skill's primary operation is API calls, making no assumptions about the user's project structure.",{"category":96,"check":97,"severity":98,"summary":99},"Trust","Issues Attention","warning","In the last 90 days, 17 issues were opened and 3 were closed, indicating a low closure rate (17.6%) and potential slow response to open issues.",{"category":101,"check":102,"severity":24,"summary":103},"Versioning","Release Management","The skill has a 'version: 1' in its frontmatter and a GitHub release tag, providing clear versioning.",{"category":105,"check":106,"severity":24,"summary":107},"Code Execution","Validation","The Python implementation for 'api_search' is expected to handle API responses gracefully and validate inputs, though explicit schema library usage is not confirmed from the provided snippets.",{"category":66,"check":109,"severity":24,"summary":110},"Unguarded Destructive Operations","This skill is read-only, performing literature searches, and thus has no destructive operations to guard.",{"category":105,"check":112,"severity":24,"summary":113},"Error Handling","The skill's 'Quality Bar' section mentions 'No crash on empty or malformed API response', indicating good error handling practices.",{"category":105,"check":115,"severity":46,"summary":116},"Logging","The skill performs read-only API calls and does not involve destructive actions or sensitive outbound calls, making extensive local audit logging not applicable.",{"category":118,"check":119,"severity":24,"summary":120},"Compliance","GDPR","The skill searches public literature databases and does not operate on personal data, thus not posing a GDPR risk.",{"category":118,"check":122,"severity":24,"summary":123},"Target market","The skill targets biomedical and life-science literature searches globally, with no specific regional limitations detected. TargetMarket is 'global'.",{"category":92,"check":125,"severity":24,"summary":126},"Runtime stability","The skill relies on standard API calls and Python, which are widely compatible, and does not appear to make OS-specific assumptions.",{"category":44,"check":128,"severity":24,"summary":129},"README","A comprehensive README file exists, detailing installation, channels, setup, and licensing, clearly stating the extension's purpose.",{"category":33,"check":131,"severity":24,"summary":132},"Tool surface size","The skill exposes a single tool ('api_search'), which is well within the ideal range.",{"category":40,"check":134,"severity":24,"summary":135},"Overlapping near-synonym tools","With only one exposed tool ('api_search'), there are no overlapping near-synonym tools.",{"category":44,"check":137,"severity":24,"summary":138},"Phantom features","All advertised capabilities, primarily PubMed search, are implemented via the 'api_search' method.",{"category":140,"check":141,"severity":24,"summary":142},"Install","Installation instruction","The README provides clear installation instructions using npm ('npx autosearch-ai') and Claude Code integration, along with verification steps.",{"category":144,"check":145,"severity":24,"summary":146},"Errors","Actionable error messages","The 'Quality Bar' section in SKILL.md explicitly states 'No crash on empty or malformed API response', implying good error handling and reporting.",{"category":148,"check":149,"severity":24,"summary":150},"Execution","Pinned dependencies","The project includes a lockfile, indicating pinned dependencies. The Python script likely uses standard libraries or has its dependencies managed.",{"category":33,"check":152,"severity":46,"summary":153},"Dry-run preview","The skill is read-only, performing literature searches, and does not have state-changing operations, making a dry-run preview not applicable.",{"category":155,"check":156,"severity":24,"summary":157},"Protocol","Idempotent retry & timeouts","The skill uses a public API which should ideally be idempotent or have retry guidance. The rate limits suggest consideration for network stability. Explicit timeouts are not detailed but expected for API clients.",{"category":118,"check":159,"severity":24,"summary":160},"Telemetry opt-in","No telemetry collection is mentioned, and the project's focus on privacy and open source suggests any telemetry would be opt-in.",{"category":40,"check":162,"severity":24,"summary":163},"Precise Purpose","The description clearly states the artifact (PubMed articles, clinical trials, medical research papers) and the user intent (biomedical and life-science literature searches).",{"category":40,"check":165,"severity":24,"summary":166},"Concise Frontmatter","The frontmatter in SKILL.md is concise, providing essential details like name, description, version, and domains without excessive length.",{"category":44,"check":168,"severity":24,"summary":169},"Concise Body","The SKILL.md body is concise, focusing on overview, when to choose, and how to search, deferring deeper material implicitly.",{"category":171,"check":172,"severity":24,"summary":173},"Context","Progressive Disclosure","The SKILL.md is concise and does not embed large amounts of material, implying deeper information would be in separate files if needed.",{"category":171,"check":175,"severity":46,"summary":176},"Forked exploration","This skill performs direct API searches and does not involve deep exploration or code review that would necessitate a forked context.",{"category":22,"check":178,"severity":24,"summary":179},"Usage examples","While explicit end-to-end examples are not shown in the provided snippets, the 'How To Search' section and the overall documentation imply clear usage patterns for literature search.",{"category":22,"check":181,"severity":24,"summary":182},"Edge cases","The 'Quality Bar' section in SKILL.md explicitly mentions handling of empty or malformed API responses, indicating consideration for edge cases.",{"category":105,"check":184,"severity":46,"summary":185},"Tool Fallback","The skill appears to directly call the PubMed API and does not rely on an external MCP server, making fallback mechanisms not applicable.",{"category":92,"check":187,"severity":24,"summary":188},"Stack assumptions","The skill is written in Python and relies on standard libraries and public APIs, with no exotic language runtimes or specific OS assumptions.",{"category":190,"check":191,"severity":24,"summary":192},"Safety","Halt on unexpected state","The 'Quality Bar' explicitly states 'No crash on empty or malformed API response', implying that unexpected states will halt the process gracefully rather than causing silent failures.",{"category":92,"check":194,"severity":24,"summary":195},"Cross-skill coupling","The skill is self-contained, focusing solely on PubMed searches, and does not appear to implicitly rely on other skills.",1778668214447,"This skill interfaces with the PubMed database to perform searches for biomedical and life-science literature, including research papers, clinical trials, and drug studies. It utilizes the E-utilities API for retrieving article IDs and metadata.",[199,200,201,202],"Searches PubMed for articles, clinical trials, and research papers","Leverages E-utilities API for efficient data retrieval","No API key required for access","Suitable for academic, medical, and literature review queries",[204,205,206,207,208],"General web search","News aggregation","Product reviews","Real-time information retrieval","Programming-related queries",[210,211],"Academic research","Medical literature review",[],"3.0.0","4.4.0","To enable AI agents to conduct precise searches within the vast PubMed database for biomedical, clinical, and life-science research literature.","The 'Issues Attention' check has a warning severity due to a low issue closure rate, which is the only warning found. All other checks passed or were not applicable.",75,"A high-quality skill for targeted biomedical literature searches.",[220,221,222,223,224],"biomedical","life-science","pubmed","research","academic","global","community",[228,229,230,231],"Researching biomedical topics and drug studies","Finding peer-reviewed medical journal articles","Locating clinical trial information","Gathering abstracts with DOI or PubMed ID for citations",[233,234,235,236],"User provides a query for biomedical or life-science literature.","Skill uses the 'api_search' method to query the PubMed E-utilities API.","API returns article IDs and metadata.","Skill returns structured search results to the user.",{"codeQuality":238,"collectedAt":240,"documentation":241,"maintenance":244,"security":250,"testCoverage":253},{"hasLockfile":239},true,1778668198328,{"descriptionLength":242,"readmeSize":243},136,5742,{"closedIssues90d":245,"forks":245,"hasChangelog":239,"manifestVersion":246,"openIssues90d":247,"pushedAt":248,"stars":249},3,"1",17,1778523503000,18,{"hasNpmPackage":251,"license":252,"smitheryVerified":251},false,"MIT",{"hasCi":239,"hasTests":239},{"updatedAt":255},1778668214554,{"basePath":257,"githubOwner":258,"githubRepo":259,"locale":18,"slug":222,"type":260},"autosearch/skills/channels/pubmed","0xmariowu","Autosearch","skill",null,{"evaluate":263,"extract":265},{"promptVersionExtension":213,"promptVersionScoring":214,"score":217,"tags":264,"targetMarket":225,"tier":226},[220,221,222,223,224],{"commitSha":266,"license":252},"HEAD",{"repoId":268},"kd745mjnht58x7nz8t1h7zqhq586n4zs",{"_creationTime":270,"_id":268,"identity":271,"providers":272,"workflow":689},1778668062084.139,{"githubOwner":258,"githubRepo":259,"sourceUrl":14},{"classify":273,"discover":663,"github":666},{"commitSha":266,"extensions":274},[275,292,304,313,321,329,337,345,353,358,366,374,385,393,401,409,417,425,433,441,449,457,465,473,481,489,497,505,513,521,529,565,573,581,589,597,605,613,621,629,637,649],{"basePath":276,"description":277,"displayName":278,"installMethods":279,"rationale":280,"selectedPaths":281,"source":290,"sourceLanguage":18,"type":291},"","Self-evolving deep research. Gets smarter every time you use it.","autosearch",{"claudeCode":12},"marketplace.json at .claude-plugin/marketplace.json",[282,285,287],{"path":283,"priority":284},".claude-plugin/marketplace.json","mandatory",{"path":286,"priority":284},"README.md",{"path":288,"priority":289},"LICENSE","high","rule","marketplace",{"basePath":276,"description":293,"displayName":278,"installMethods":294,"license":252,"rationale":295,"selectedPaths":296,"source":290,"sourceLanguage":18,"type":303},"Self-evolving deep research. Gets smarter every time you use it. Searches across channels, synthesizes cited reports, and learns which queries and platforms work best.",{"claudeCode":278},"plugin manifest at .claude-plugin/plugin.json",[297,299,300,301],{"path":298,"priority":284},".claude-plugin/plugin.json",{"path":286,"priority":284},{"path":288,"priority":289},{"path":302,"priority":289},"commands/autosearch.md","plugin",{"basePath":305,"description":306,"displayName":307,"installMethods":308,"rationale":309,"selectedPaths":310,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/channels/ddgs","DuckDuckGo Search — free general web search with no auth, use as broad default for any English or mixed query.","ddgs",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/channels/ddgs/SKILL.md",[311],{"path":312,"priority":284},"SKILL.md",{"basePath":314,"description":315,"displayName":316,"installMethods":317,"rationale":318,"selectedPaths":319,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/channels/dockerhub","Use for Docker image and container registry searches when query involves Docker images, containers, or deployment artifacts.","dockerhub",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/channels/dockerhub/SKILL.md",[320],{"path":312,"priority":284},{"basePath":322,"description":323,"displayName":324,"installMethods":325,"rationale":326,"selectedPaths":327,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/channels/instagram","Instagram public posts and reels search via TikHub. Best for product demos, tutorials, and visual content. English-first but works with any language keyword.","instagram",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/channels/instagram/SKILL.md",[328],{"path":312,"priority":284},{"basePath":330,"description":331,"displayName":332,"installMethods":333,"rationale":334,"selectedPaths":335,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/channels/kr36","Chinese tech business news, startup funding, and industry analysis from 36kr.","kr36",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/channels/kr36/SKILL.md",[336],{"path":312,"priority":284},{"basePath":338,"description":339,"displayName":340,"installMethods":341,"rationale":342,"selectedPaths":343,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/channels/linkedin","LinkedIn public pages — company profiles, job posts, professional articles via Jina Reader (no auth needed for public content).","linkedin",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/channels/linkedin/SKILL.md",[344],{"path":312,"priority":284},{"basePath":346,"description":347,"displayName":348,"installMethods":349,"rationale":350,"selectedPaths":351,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/channels/openalex","Search scholarly works through OpenAlex's public works search API with open-access URL fallback.","openalex",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/channels/openalex/SKILL.md",[352],{"path":312,"priority":284},{"basePath":257,"description":10,"displayName":222,"installMethods":354,"rationale":355,"selectedPaths":356,"source":290,"sourceLanguage":18,"type":260},{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/channels/pubmed/SKILL.md",[357],{"path":312,"priority":284},{"basePath":359,"description":360,"displayName":361,"installMethods":362,"rationale":363,"selectedPaths":364,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/channels/reddit","Reddit community discussions, user experience reports, and topic debates via the public search.json endpoint.","reddit",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/channels/reddit/SKILL.md",[365],{"path":312,"priority":284},{"basePath":367,"description":368,"displayName":369,"installMethods":370,"rationale":371,"selectedPaths":372,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/channels/searxng","Use for general web meta-search via a local SearXNG instance when broad multi-engine web coverage is needed and SEARXNG_URL is configured.","searxng",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/channels/searxng/SKILL.md",[373],{"path":312,"priority":284},{"basePath":375,"description":376,"displayName":377,"installMethods":378,"rationale":379,"selectedPaths":380,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/channels/stackoverflow","Programming Q&A with community-voted answers across 200+ technical tags via api.stackexchange.com.","stackoverflow",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/channels/stackoverflow/SKILL.md",[381,382],{"path":312,"priority":284},{"path":383,"priority":384},"experience.md","medium",{"basePath":386,"description":387,"displayName":388,"installMethods":389,"rationale":390,"selectedPaths":391,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/channels/tieba","Use for Chinese community forum discussions on Baidu Tieba when query involves Chinese tech topics, product feedback, or hobby communities not covered by xiaohongshu/zhihu.","tieba",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/channels/tieba/SKILL.md",[392],{"path":312,"priority":284},{"basePath":394,"description":395,"displayName":396,"installMethods":397,"rationale":398,"selectedPaths":399,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/channels/v2ex","Chinese developer community discussions on programming, tech, and career — useful for native-zh developer sentiment and niche technical troubleshooting.","v2ex",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/channels/v2ex/SKILL.md",[400],{"path":312,"priority":284},{"basePath":402,"description":403,"displayName":404,"installMethods":405,"rationale":406,"selectedPaths":407,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/channels/wikidata","Structured entity data (people, places, concepts) from Wikidata knowledge graph.","wikidata",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/channels/wikidata/SKILL.md",[408],{"path":312,"priority":284},{"basePath":410,"description":411,"displayName":412,"installMethods":413,"rationale":414,"selectedPaths":415,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/channels/wikipedia","Authoritative encyclopedia articles via the Wikipedia Action API (English edition).","wikipedia",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/channels/wikipedia/SKILL.md",[416],{"path":312,"priority":284},{"basePath":418,"description":419,"displayName":420,"installMethods":421,"rationale":422,"selectedPaths":423,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/channels/xueqiu","Chinese stock market platform — stock search, trending posts, hot stock rankings. Use for financial/investment queries in Chinese.","xueqiu",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/channels/xueqiu/SKILL.md",[424],{"path":312,"priority":284},{"basePath":426,"description":427,"displayName":428,"installMethods":429,"rationale":430,"selectedPaths":431,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/meta/channel-selection","Group-first channel selection algorithm for v2 tool-supplier architecture. Given a research query + clarify rubrics + channel_priority hints, picks 1-3 relevant groups from the router index, then 3-8 leaf channels from within those groups. Replaces flat-rank selection across 41 channels with a two-stage pick so runtime AI never has to read all 41 SKILL.md bodies.","channel-selection",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/meta/channel-selection/SKILL.md",[432],{"path":312,"priority":284},{"basePath":434,"description":435,"displayName":436,"installMethods":437,"rationale":438,"selectedPaths":439,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/meta/citation-index","Deduplicate URLs across all sources, assign stable citation numbers, and merge citations from multiple subagents / sections into one consistent reference list. Prevents \"same URL cited as [3] in one paragraph and [17] in another\" and \"different URLs merged under [5]\" bugs that come from per-section synthesis.","citation-index",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/meta/citation-index/SKILL.md",[440],{"path":312,"priority":284},{"basePath":442,"description":443,"displayName":444,"installMethods":445,"rationale":446,"selectedPaths":447,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/meta/context-retention-policy","Session-level policy for keeping the runtime AI's context window healthy across long research — keep-last-k tool results, offload older evidence to disk, trigger compaction at thresholds. Borrows MiroThinker's keep_tool_result, deepagents' summarization middleware, and deer-flow's SummarizationEvent pattern. Orthogonal to assemble-context (which is per-synthesis-step); this is per-session.","context-retention-policy",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/meta/context-retention-policy/SKILL.md",[448],{"path":312,"priority":284},{"basePath":450,"description":451,"displayName":452,"installMethods":453,"rationale":454,"selectedPaths":455,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/meta/delegate-subtask","Define the execution contract for isolating a research sub-task — input schema, budget, return summary, evidence list, failure status. Complements decompose-task (which only splits the problem) by giving each split a bounded, auditable execution unit the runtime AI can farm out to a sub-agent or parallel session.","delegate-subtask",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/meta/delegate-subtask/SKILL.md",[456],{"path":312,"priority":284},{"basePath":458,"description":459,"displayName":460,"installMethods":461,"rationale":462,"selectedPaths":463,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/meta/experience-capture","Append a single skill-execution event to the per-skill `experience/patterns.jsonl` file. Runs every time a leaf skill is used. Append-only — runtime AI never reads this file directly; only the compacted `experience.md` digest. Fast tier, no LLM required.","experience-capture",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/meta/experience-capture/SKILL.md",[464],{"path":312,"priority":284},{"basePath":466,"description":467,"displayName":468,"installMethods":469,"rationale":470,"selectedPaths":471,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/meta/experience-compact","Promote recurring patterns from `experience/patterns.jsonl` into the compact `experience.md` digest (≤120 lines, read by runtime AI before calling the skill). Triggers on N-events / file-size / user-feedback / session-end. Guards against single-success noise and pollution via promotion thresholds.","experience-compact",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/meta/experience-compact/SKILL.md",[472],{"path":312,"priority":284},{"basePath":474,"description":475,"displayName":476,"installMethods":477,"rationale":478,"selectedPaths":479,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/meta/graph-search-plan","Represent a research plan as a directed graph — nodes are sub-questions, edges are \"answer-depends-on\" dependencies. Independent nodes can be executed in parallel; downstream nodes wait for their predecessors. Borrowed from MindSearch's WebSearchGraph pattern. Replaces list/tree decomposition for plans with non-linear dependencies.","graph-search-plan",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/meta/graph-search-plan/SKILL.md",[480],{"path":312,"priority":284},{"basePath":482,"description":483,"displayName":484,"installMethods":485,"rationale":486,"selectedPaths":487,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/meta/model-routing","Advisory skill — three-tier (Fast / Standard / Best) model routing catalog for autosearch skills. Tells the runtime AI which tier each leaf skill needs, and how to escalate or de-escalate. Autosearch does not switch models itself; the runtime AI is the decision-maker.","model-routing",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/meta/model-routing/SKILL.md",[488],{"path":312,"priority":284},{"basePath":490,"description":491,"displayName":492,"installMethods":493,"rationale":494,"selectedPaths":495,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/meta/perspective-questioning","Generate questions from 3-6 stakeholder / expert personas to expand coverage before searching. Complements decompose-task (which splits by task structure) by splitting by viewpoint. Borrowed from STORM's pre-writing research pattern — asking \"what would an investor / a security researcher / an end-user / a maintainer\" each want to know.","perspective-questioning",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/meta/perspective-questioning/SKILL.md",[496],{"path":312,"priority":284},{"basePath":498,"description":499,"displayName":500,"installMethods":501,"rationale":502,"selectedPaths":503,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/meta/recent-signal-fusion","Unify signals from multiple recency-sensitive channels (Reddit, X, Hacker News, Weibo, YouTube, GitHub activity, Polymarket, etc.) into a single time-weighted candidate list. Clusters semantic near-duplicates across sources, weights by recency + platform reliability, and produces a ranked \"what's happening recently\" bundle for the runtime AI to synthesize.","recent-signal-fusion",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/meta/recent-signal-fusion/SKILL.md",[504],{"path":312,"priority":284},{"basePath":506,"description":507,"displayName":508,"installMethods":509,"rationale":510,"selectedPaths":511,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/meta/reflective-search-loop","Explicit loop state for multi-round research — maintains gaps, visited URLs, bad URLs, failed-evaluator feedback, and decides whether to fire another round of queries or stop. Inspired by WebThinker's in-band search protocol, node-deepresearch's agent.ts state, and Scira's extreme-search.","reflective-search-loop",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/meta/reflective-search-loop/SKILL.md",[512],{"path":312,"priority":284},{"basePath":514,"description":515,"displayName":516,"installMethods":517,"rationale":518,"selectedPaths":519,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/meta/tikhub-fallback","Decision tree for when to escalate from free native Chinese channels (bilibili / weibo / xiaohongshu / douyin / zhihu) to the paid TikHub fallback. ~$0.0036 per request; 5 platforms covered. Tells the runtime AI when the cost is worth paying vs. when the free paths are enough.","tikhub-fallback",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/meta/tikhub-fallback/SKILL.md",[520],{"path":312,"priority":284},{"basePath":522,"description":523,"displayName":524,"installMethods":525,"rationale":526,"selectedPaths":527,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/meta/trace-harvest","Distill reusable knowledge from successful session tool-call traces. Reads runtime tool-call logs + Evidence outputs + user acceptance signals, filters successful paths, and writes compact patterns to per-skill experience/patterns.jsonl. Differs from outcome-tracker (which only records downstream acceptance) by parsing the full trace structure — inputs, tool choices, branches, recoveries.","trace-harvest",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/meta/trace-harvest/SKILL.md",[528],{"path":312,"priority":284},{"basePath":530,"description":531,"displayName":532,"installMethods":533,"rationale":534,"selectedPaths":535,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/router","Route an autosearch task to the smallest relevant group of skills, then load only the matching leaf skills. First line of defense against loading all 80+ skill SKILL.md files at session start.","router",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/router/SKILL.md",[536,537,539,541,543,545,547,549,551,553,555,557,559,561,563],{"path":312,"priority":284},{"path":538,"priority":384},"references/groups/channels-academic.md",{"path":540,"priority":384},"references/groups/channels-chinese-ugc.md",{"path":542,"priority":384},"references/groups/channels-cn-tech.md",{"path":544,"priority":384},"references/groups/channels-code-package.md",{"path":546,"priority":384},"references/groups/channels-community-en.md",{"path":548,"priority":384},"references/groups/channels-generic-web.md",{"path":550,"priority":384},"references/groups/channels-market-product.md",{"path":552,"priority":384},"references/groups/channels-social-career.md",{"path":554,"priority":384},"references/groups/channels-video-audio.md",{"path":556,"priority":384},"references/groups/tools-fetch-render.md",{"path":558,"priority":384},"references/groups/workflow-growth.md",{"path":560,"priority":384},"references/groups/workflow-planning.md",{"path":562,"priority":384},"references/groups/workflow-quality.md",{"path":564,"priority":384},"references/groups/workflow-synthesis.md",{"basePath":566,"description":567,"displayName":568,"installMethods":569,"rationale":570,"selectedPaths":571,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/tools/fetch-crawl4ai","Deep URL fetch using crawl4ai (Playwright-powered) for JS-rendered pages, anti-bot sites, and dynamic content. Slower than fetch-jina but handles sites that block simple fetchers. Requires user-installed crawl4ai package.","fetch-crawl4ai",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/tools/fetch-crawl4ai/SKILL.md",[572],{"path":312,"priority":284},{"basePath":574,"description":575,"displayName":576,"installMethods":577,"rationale":578,"selectedPaths":579,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/tools/fetch-firecrawl","Use to scrape a URL into clean Markdown when the page is JS-rendered, behind anti-bot, or a PDF — and FIRECRAWL_API_KEY is available. Degrades gracefully to warn when key is missing.","fetch-firecrawl",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/tools/fetch-firecrawl/SKILL.md",[580],{"path":312,"priority":284},{"basePath":582,"description":583,"displayName":584,"installMethods":585,"rationale":586,"selectedPaths":587,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/tools/fetch-jina","Fetch any public URL as clean Markdown via Jina Reader (r.jina.ai). Best for articles, docs, blog posts; weak on heavy anti-bot sites (Zhihu, Xiaohongshu).","fetch-jina",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/tools/fetch-jina/SKILL.md",[588],{"path":312,"priority":284},{"basePath":590,"description":591,"displayName":592,"installMethods":593,"rationale":594,"selectedPaths":595,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/tools/fetch-playwright","Interactive browser automation (click, type, wait, screenshot, navigate) for dynamic pages that need user-like actions — pagination, login, form submission, infinite scroll. Uses Microsoft's official @playwright/mcp server; no autosearch-side Python. Use when fetch-jina and fetch-crawl4ai still cannot get the content.","fetch-playwright",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/tools/fetch-playwright/SKILL.md",[596],{"path":312,"priority":284},{"basePath":598,"description":599,"displayName":600,"installMethods":601,"rationale":602,"selectedPaths":603,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/tools/mcporter","Route the runtime AI to free third-party MCP servers via the upstream `mcporter` router — semantic web search (Exa), Chinese UGC (Weibo / Douyin / Xiaohongshu), and professional (LinkedIn). Zero API keys; free fallback path before reaching for paid TikHub on hard Chinese anti-bot platforms.","mcporter",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/tools/mcporter/SKILL.md",[604],{"path":312,"priority":284},{"basePath":606,"description":607,"displayName":608,"installMethods":609,"rationale":610,"selectedPaths":611,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/tools/video-to-text-bcut","Transcribe video/audio URL to text + word-level timestamps using Bilibili Bcut ASR API (free, no API key). Preferred for Chinese content — Bcut gives character-level timestamps vs Whisper word-level. Returns text + segments [{start, end, text}]. Requires yt-dlp + ffmpeg.","video-to-text-bcut",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/tools/video-to-text-bcut/SKILL.md",[612],{"path":312,"priority":284},{"basePath":614,"description":615,"displayName":616,"installMethods":617,"rationale":618,"selectedPaths":619,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/tools/video-to-text-groq","Transcribe video/audio URL or local file to text + SRT using yt-dlp + Groq Whisper API (free tier). Preferred default for v2 transcription. Returns raw text and segments; summary is caller's responsibility.","video-to-text-groq",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/tools/video-to-text-groq/SKILL.md",[620],{"path":312,"priority":284},{"basePath":622,"description":623,"displayName":624,"installMethods":625,"rationale":626,"selectedPaths":627,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/tools/video-to-text-local","Transcribe video/audio URL or local file to text + SRT using yt-dlp + local mlx-whisper (Apple Silicon). Free, offline, fastest on M-series Macs. Opt-in advanced path for users with Apple Silicon + mlx-whisper installed. Returns raw text and segments; summary is caller's responsibility.","video-to-text-local",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/tools/video-to-text-local/SKILL.md",[628],{"path":312,"priority":284},{"basePath":630,"description":631,"displayName":632,"installMethods":633,"rationale":634,"selectedPaths":635,"source":290,"sourceLanguage":18,"type":260},"autosearch/skills/tools/video-to-text-openai","Transcribe video/audio URL or local file to text + SRT using yt-dlp + OpenAI Whisper API. Paid fallback for v2 transcription when Groq is rate-limited or unavailable. Returns raw text and segments; summary is caller's responsibility.","video-to-text-openai",{"claudeCode":12},"SKILL.md frontmatter at autosearch/skills/tools/video-to-text-openai/SKILL.md",[636],{"path":312,"priority":284},{"basePath":638,"description":639,"displayName":640,"installMethods":641,"license":252,"rationale":642,"selectedPaths":643,"source":290,"sourceLanguage":18,"type":648},"npm","AutoSearch — deep research MCP server for AI developers","autosearch-ai",{"npm":640},"cli ecosystem detected at npm",[644,646],{"path":645,"priority":284},"package.json",{"path":647,"priority":384},"bin/autosearch-ai.js","cli",{"basePath":276,"displayName":278,"installMethods":650,"rationale":651,"selectedPaths":652,"source":290,"sourceLanguage":18,"type":662},{"pypi":278},"pyproject.toml with mcp/fastmcp dependency + scripts at pyproject.toml",[653,654,656,657,658,660],{"path":645,"priority":284},{"path":655,"priority":284},"pyproject.toml",{"path":286,"priority":284},{"path":288,"priority":289},{"path":659,"priority":384},"autosearch/cli/main.py",{"path":661,"priority":384},"autosearch/mcp/cli.py","mcp",{"sources":664},[665],"manual",{"closedIssues90d":245,"description":667,"forks":245,"homepage":668,"license":252,"openIssues90d":247,"pushedAt":248,"readmeSize":243,"stars":249,"topics":669},"Open-source deep research for AI agents: 40 channels, 10+ Chinese sources.","https://www.autosearch.dev",[670,671,672,673,674,675,676,677,678,679,680,681,682,683,684,685,686,662,687,688],"claude-code-plugin","deep-research","agentic-search","chinese-search","mcp-server","research-agent","agent-infrastructure","open-source","ai-agent","ai-agents","autogen","claude-code","cline","cursor","gpt-researcher","langchain","llamaindex","perplexity-alternative","tavily-alternative",{"classifiedAt":690,"discoverAt":691,"extractAt":692,"githubAt":692,"updatedAt":690},1778668067697,1778668062084,1778668065669,[224,220,221,222,223],{"evaluatedAt":255,"extractAt":695,"updatedAt":255},1778668068135,[],[698,728,757,785,807,835],{"_creationTime":699,"_id":700,"community":701,"display":702,"identity":708,"providers":713,"relations":722,"tags":724,"workflow":725},1778691799740.48,"k170wty24x8dpw3h33f885hj4986nwnc",{"reviewCount":8},{"description":703,"installMethods":704,"name":706,"sourceUrl":707},"Conduct comprehensive, systematic literature reviews using multiple academic databases (PubMed, arXiv, bioRxiv, Semantic Scholar, etc.). This skill should be used when conducting systematic literature reviews, meta-analyses, research synthesis, or comprehensive literature searches across biomedical, scientific, and technical domains. Creates professionally formatted markdown documents and PDFs with verified citations in multiple citation styles (APA, Nature, Vancouver, etc.).",{"claudeCode":705},"K-Dense-AI/claude-scientific-skills","Literature Review","https://github.com/K-Dense-AI/claude-scientific-skills",{"basePath":709,"githubOwner":710,"githubRepo":711,"locale":18,"slug":712,"type":260},"scientific-skills/literature-review","K-Dense-AI","claude-scientific-skills","literature-review",{"evaluate":714,"extract":721},{"promptVersionExtension":213,"promptVersionScoring":214,"score":715,"tags":716,"targetMarket":225,"tier":720},100,[712,223,224,717,718,719],"documentation","pdf-generation","citation-management","verified",{"commitSha":266,"license":252},{"repoId":723},"kd79rphh5gexy91xmpxc05h5mh86mm9r",[224,719,717,712,718,223],{"evaluatedAt":726,"extractAt":727,"updatedAt":726},1778693072547,1778691799740,{"_creationTime":729,"_id":730,"community":731,"display":732,"identity":738,"providers":743,"relations":751,"tags":753,"workflow":754},1778675145461.8677,"k17fng4atqmrf5j27jnkhq15rd86ns0j",{"reviewCount":8},{"description":733,"installMethods":734,"name":736,"sourceUrl":737},"Query Open Targets Platform for target-disease associations, drug target discovery, tractability/safety data, genetics/omics evidence, known drugs, for therapeutic target identification. Part of the AlterLab Academic Skills suite.",{"claudeCode":735},"AlterLab-IEU/AlterLab-Academic-Skills","Open Targets Platform Query Skill","https://github.com/AlterLab-IEU/AlterLab-Academic-Skills",{"basePath":739,"githubOwner":740,"githubRepo":741,"locale":18,"slug":742,"type":260},"skills/databases/alterlab-opentargets","AlterLab-IEU","AlterLab-Academic-Skills","alterlab-opentargets",{"evaluate":744,"extract":750},{"promptVersionExtension":213,"promptVersionScoring":214,"score":715,"tags":745,"targetMarket":225,"tier":720},[220,746,747,748,223,749],"drug-discovery","genomics","pharmacology","opentargets",{"commitSha":266,"license":252},{"repoId":752},"kd7fqvj70pvyn4r3q9kctpnd7d86mfqd",[220,746,747,749,748,223],{"evaluatedAt":755,"extractAt":756,"updatedAt":755},1778677740053,1778675145461,{"_creationTime":758,"_id":759,"community":760,"display":761,"identity":767,"providers":771,"relations":778,"tags":781,"workflow":782},1778695548458.4038,"k179pva4fdcsabphehnng6njts86mb8k",{"reviewCount":8},{"description":762,"installMethods":763,"name":765,"sourceUrl":766},"Survey and synthesize theoretical literature on a specific topic, identifying seminal papers, key results, open problems, and cross-domain connections. Use when starting research on an unfamiliar theoretical topic, writing a literature review for a paper or thesis, identifying open problems and research gaps, finding cross-domain connections, or evaluating the novelty of a proposed theoretical contribution against existing work.\n",{"claudeCode":764},"pjt222/agent-almanac","survey-theoretical-literature","https://github.com/pjt222/agent-almanac",{"basePath":768,"githubOwner":769,"githubRepo":770,"locale":18,"slug":765,"type":260},"skills/survey-theoretical-literature","pjt222","agent-almanac",{"evaluate":772,"extract":777},{"promptVersionExtension":213,"promptVersionScoring":214,"score":773,"tags":774,"targetMarket":225,"tier":720},99,[223,712,775,776,224],"synthesis","theoretical-science",{"commitSha":266},{"parentExtensionId":779,"repoId":780},"k170h0janaa9kwn7cfgfz2ykss86mmh9","kd7aryv63z61j39n2td1aeqkvh86mh12",[224,712,223,775,776],{"evaluatedAt":783,"extractAt":784,"updatedAt":783},1778701846608,1778695548458,{"_creationTime":786,"_id":787,"community":788,"display":789,"identity":793,"providers":795,"relations":803,"tags":804,"workflow":805},1778675145461.8684,"k1796e7tzwnvf2rpg1kpyr59j186nnnk",{"reviewCount":8},{"description":790,"installMethods":791,"name":792,"sourceUrl":737},"Direct REST API access to PubMed. Advanced Boolean/MeSH queries, E-utilities API, batch processing, citation management. For Python workflows, prefer biopython (Bio.Entrez). Use this for direct HTTP/REST work or custom API implementations. Part of the AlterLab Academic Skills suite.",{"claudeCode":735},"alterlab-pubmed",{"basePath":794,"githubOwner":740,"githubRepo":741,"locale":18,"slug":792,"type":260},"skills/databases/alterlab-pubmed",{"evaluate":796,"extract":802},{"promptVersionExtension":213,"promptVersionScoring":214,"score":797,"tags":798,"targetMarket":225,"tier":720},95,[222,220,799,800,801,223],"literature-search","api","eutilities",{"commitSha":266},{"repoId":752},[800,220,801,799,222,223],{"evaluatedAt":806,"extractAt":756,"updatedAt":806},1778677821246,{"_creationTime":808,"_id":809,"community":810,"display":811,"identity":817,"providers":821,"relations":828,"tags":831,"workflow":832},1778699234184.6135,"k175frmf44tn80mcd6gvw1c1th86ngq9",{"reviewCount":8},{"description":812,"installMethods":813,"name":815,"sourceUrl":816},"Invoke parallel document-specialist agents for external web searches and documentation lookup",{"claudeCode":814},"Yeachan-Heo/oh-my-claudecode","external-context","https://github.com/Yeachan-Heo/oh-my-claudecode",{"basePath":818,"githubOwner":819,"githubRepo":820,"locale":18,"slug":815,"type":260},"skills/external-context","Yeachan-Heo","oh-my-claudecode",{"evaluate":822,"extract":827},{"promptVersionExtension":213,"promptVersionScoring":214,"score":715,"tags":823,"targetMarket":225,"tier":720},[824,717,223,825,826],"search","information-retrieval","multi-agent",{"commitSha":266},{"parentExtensionId":829,"repoId":830},"k17brg5egdw1jbncj1j4wfv3fh86n639","kd74zv63fryf9prygtq7gf4es986n22y",[717,825,826,223,824],{"evaluatedAt":833,"extractAt":834,"updatedAt":833},1778699449790,1778699234184,{"_creationTime":836,"_id":837,"community":838,"display":839,"identity":845,"providers":850,"relations":858,"tags":861,"workflow":862},1778695116697.199,"k17cex5hqwje7qgvts5evct1d186nd4m",{"reviewCount":8},{"description":840,"installMethods":841,"name":843,"sourceUrl":844},"Records research provenance as a post-task epilogue, scanning conversation history at the end of a coding or research session to extract decisions, experiments, dead ends, claims, heuristics, and pivots, and writing them into the ara/ directory with user-vs-AI provenance tags. Use as a session epilogue — never during execution — to maintain a faithful, auditable trace of how a research project actually evolved.",{"claudeCode":842},"Orchestra-Research/AI-Research-SKILLs","ARA Research Manager","https://github.com/Orchestra-Research/AI-Research-SKILLs",{"basePath":846,"githubOwner":847,"githubRepo":848,"locale":18,"slug":849,"type":260},"22-agent-native-research-artifact/research-manager","Orchestra-Research","AI-Research-SKILLs","research-manager",{"evaluate":851,"extract":857},{"promptVersionExtension":213,"promptVersionScoring":214,"score":715,"tags":852,"targetMarket":225,"tier":720},[223,853,854,855,856],"provenance","knowledge-management","session-logging","ara",{"commitSha":266,"license":252},{"parentExtensionId":859,"repoId":860},"k17155ws9qc0hw7a568bg79sfd86max8","kd70hj1y80mhra5xm5g188j5n586mg18",[856,854,853,223,855],{"evaluatedAt":863,"extractAt":864,"updatedAt":863},1778697541177,1778695116697]