{"id":"b1cd8d56-3f16-4e2b-ae55-a469e9998af2","shortId":"KNTGq8","kind":"skill","title":"agentic-search","tagline":"Grok-primary deep research skill for source-backed web work. Use when the task needs current web research, grounded citations, supplementary Tavily/Firecrawl source discovery, high-fidelity page-to-Markdown fetch, Tavily site mapping, verbatim quote extraction, source reranking, ","description":"# agentic-search\n\n| # | Script | Intent |\n|---|---|---|\n| 1 | `scripts/agentic_search.py` | Research a topic — AI-reasoned answer with citations |\n| 2 | `scripts/agentic_fetch.py` | Full page → Markdown, no summarization |\n| 3 | `scripts/agentic_map.py` | Enumerate URLs under a site (Tavily-only) |\n| 4 | `scripts/agentic_extract.py` | Verbatim title + 2–4 quotes from a URL |\n| 5 | `scripts/agentic_rank.py` | Rerank session sources by a refined query |\n| 6 | `scripts/agentic_get_sources.py` | Retrieve or list cached sessions |\n\n## How to invoke\n\nAll scripts use PEP 723 inline deps — always invoke via `uv run`. Allow up to 3 minutes per call.\n\n```bash\nuv run scripts/agentic_search.py --query \"...\" [--extra-sources N] [--auto-fetch-top N] [--platform \"...\"]\nuv run scripts/agentic_fetch.py --url \"...\" [--engine auto|tavily|firecrawl|grok]\nuv run scripts/agentic_map.py --url \"...\" [--instructions \"...\"] [--limit N]\nuv run scripts/agentic_extract.py --url \"...\" [--session-id S]\nuv run scripts/agentic_rank.py --query \"...\" --session-id S\nuv run scripts/agentic_get_sources.py --list | --session-id S\n```\n\n## Session workflow\n\n`agentic_search` generates a `session_id` persisted to disk — the connective tissue for composing steps:\n\n```\nagentic_search → session_id\n     ├─→ agentic_rank --session-id S --query Q     (rerank in place; mutates session)\n     ├─→ agentic_extract --session-id S --url U    (append verbatim quotes)\n     └─→ agentic_get_sources --session-id S        (inspect full session)\n```\n\nSessions survive between invocations but may be cleared on OS reboot.\n\n## Operating discipline\n\nLoad the relevant reference before composing a non-trivial query — don't load all four upfront.\n\n- **`references/search-discipline.md`** — two-pass methodology, citation contract, time-context heuristic. Read for research tasks.\n- **`references/fetch-fidelity.md`** — engine trade-offs, fidelity guarantees, when not to fetch. Read before presenting extracted content.\n- **`references/extract-and-rank.md`** — extract vs fetch, when to rank, session composition patterns. Read for multi-step workflows.\n- **`references/provider-quirks.md`** — env vars, provider schemas, retry policy, debugging. Read when configuring or debugging.\n\n## Failure modes\n\n- No `GROK_API_KEY` / `GROK_API_URL` → config error on any Grok-using script.\n- `agentic_fetch --engine auto` total failure → retry with `--engine grok` or fall back to built-in `WebFetch`.\n- `agentic_map` without `TAVILY_API_KEY` → hard exit.\n- `agentic_rank --session-id S` session expired → re-run `agentic_search`.\n- `sources_count: 0` with non-empty `content` → Grok used training data; check `providers_used` (`grok-web-search` = real citations, `grok` = heuristic fallback).","tags":["agentic","search","webmaton","appautomaton","agent-skills","ai-agents","browser-automation","claude-code","codex","html-to-markdown","playwright","web-automation"],"capabilities":["skill","source-appautomaton","skill-agentic-search","topic-agent-skills","topic-ai-agents","topic-browser-automation","topic-claude-code","topic-codex","topic-html-to-markdown","topic-playwright","topic-web-automation","topic-web-scraping"],"categories":["webmaton"],"synonyms":[],"warnings":[],"endpointUrl":"https://skills.sh/appautomaton/webmaton/agentic-search","protocol":"skill","transport":"skills-sh","auth":{"type":"none","details":{"cli":"npx skills add appautomaton/webmaton","source_repo":"https://github.com/appautomaton/webmaton","install_from":"skills.sh"}},"qualityScore":"0.458","qualityRationale":"deterministic score 0.46 from registry signals: · indexed on github topic:agent-skills · 17 github stars · SKILL.md body (2,742 chars)","verified":false,"liveness":"unknown","lastLivenessCheck":null,"agentReviews":{"count":0,"score_avg":null,"cost_usd_avg":null,"success_rate":null,"latency_p50_ms":null,"narrative_summary":null,"summary_updated_at":null},"enrichmentModel":"deterministic:skill-github:v1","enrichmentVersion":1,"enrichedAt":"2026-05-18T19:06:02.127Z","embedding":null,"createdAt":"2026-05-06T01:03:38.329Z","updatedAt":"2026-05-18T19:06:02.127Z","lastSeenAt":"2026-05-18T19:06:02.127Z","tsv":"'0':385 '1':51 '2':62,83 '3':69,123 '4':79,84 '5':89 '6':98 '723':112 'agent':2,47,184,199,203,216,227,344,362,370,381 'agentic-search':1,46 'ai':57 'ai-reason':56 'allow':120 'alway':115 'answer':59 'api':331,334,366 'append':224 'auto':137,147,347 'auto-fetch-top':136 'back':13,356 'bash':127 'built':359 'built-in':358 'cach':103 'call':126 'check':395 'citat':25,61,272,403 'clear':244 'compos':197,255 'composit':306 'config':336 'configur':324 'connect':194 'content':297,390 'context':276 'contract':273 'count':384 'current':21 'data':394 'debug':321,326 'deep':7 'dep':114 'disciplin':249 'discoveri':29 'disk':192 'empti':389 'engin':146,283,346,352 'enumer':71 'env':315 'error':337 'exit':369 'expir':377 'extra':133 'extra-sourc':132 'extract':43,217,296,299 'failur':327,349 'fall':355 'fallback':406 'fetch':37,138,292,301,345 'fidel':32,287 'firecrawl':149 'four':265 'full':64,235 'generat':186 'get':228 'grok':5,150,330,333,341,353,391,399,404 'grok-primari':4 'grok-us':340 'grok-web-search':398 'ground':24 'guarante':288 'hard':368 'heurist':277,405 'high':31 'high-fidel':30 'id':164,172,180,189,202,207,220,232,374 'inlin':113 'inspect':234 'instruct':155 'intent':50 'invoc':240 'invok':107,116 'key':332,367 'limit':156 'list':102,177 'load':250,263 'map':40,363 'markdown':36,66 'may':242 'methodolog':271 'minut':124 'mode':328 'multi':311 'multi-step':310 'mutat':214 'n':135,140,157 'need':20 'non':258,388 'non-empti':387 'non-trivi':257 'off':286 'oper':248 'os':246 'page':34,65 'page-to-markdown':33 'pass':270 'pattern':307 'pep':111 'per':125 'persist':190 'place':213 'platform':141 'polici':320 'present':295 'primari':6 'provid':317,396 'q':210 'queri':97,131,169,209,260 'quot':42,85,226 'rank':204,304,371 're':379 're-run':378 'read':278,293,308,322 'real':402 'reason':58 'reboot':247 'refer':253 'references/extract-and-rank.md':298 'references/fetch-fidelity.md':282 'references/provider-quirks.md':314 'references/search-discipline.md':267 'refin':96 'relev':252 'rerank':45,91,211 'research':8,23,53,280 'retri':319,350 'retriev':100 'run':119,129,143,152,159,167,175,380 'schema':318 'script':49,109,343 'scripts/agentic_extract.py':80,160 'scripts/agentic_fetch.py':63,144 'scripts/agentic_get_sources.py':99,176 'scripts/agentic_map.py':70,153 'scripts/agentic_rank.py':90,168 'scripts/agentic_search.py':52,130 'search':3,48,185,200,382,401 'session':92,104,163,171,179,182,188,201,206,215,219,231,236,237,305,373,376 'session-id':162,170,178,205,218,230,372 'site':39,75 'skill':9 'skill-agentic-search' 'sourc':12,28,44,93,134,229,383 'source-appautomaton' 'source-back':11 'step':198,312 'summar':68 'supplementari':26 'surviv':238 'task':19,281 'tavili':38,77,148,365 'tavily-on':76 'tavily/firecrawl':27 'time':275 'time-context':274 'tissu':195 'titl':82 'top':139 'topic':55 'topic-agent-skills' 'topic-ai-agents' 'topic-browser-automation' 'topic-claude-code' 'topic-codex' 'topic-html-to-markdown' 'topic-playwright' 'topic-web-automation' 'topic-web-scraping' 'total':348 'trade':285 'trade-off':284 'train':393 'trivial':259 'two':269 'two-pass':268 'u':223 'upfront':266 'url':72,88,145,154,161,222,335 'use':16,110,342,392,397 'uv':118,128,142,151,158,166,174 'var':316 'verbatim':41,81,225 'via':117 'vs':300 'web':14,22,400 'webfetch':361 'without':364 'work':15 'workflow':183,313","prices":[{"id":"20152801-37a9-4ee5-9ca1-48178b405853","listingId":"b1cd8d56-3f16-4e2b-ae55-a469e9998af2","amountUsd":"0","unit":"free","nativeCurrency":null,"nativeAmount":null,"chain":null,"payTo":null,"paymentMethod":"skill-free","isPrimary":true,"details":{"org":"appautomaton","category":"webmaton","install_from":"skills.sh"},"createdAt":"2026-05-06T01:03:38.329Z"}],"sources":[{"listingId":"b1cd8d56-3f16-4e2b-ae55-a469e9998af2","source":"github","sourceId":"appautomaton/webmaton/agentic-search","sourceUrl":"https://github.com/appautomaton/webmaton/tree/main/skills/agentic-search","isPrimary":false,"firstSeenAt":"2026-05-06T01:03:38.329Z","lastSeenAt":"2026-05-18T19:06:02.127Z"}],"details":{"listingId":"b1cd8d56-3f16-4e2b-ae55-a469e9998af2","quickStartSnippet":null,"exampleRequest":null,"exampleResponse":null,"schema":null,"openapiUrl":null,"agentsTxtUrl":null,"citations":[],"useCases":[],"bestFor":[],"notFor":[],"kindDetails":{"org":"appautomaton","slug":"agentic-search","github":{"repo":"appautomaton/webmaton","stars":17,"topics":["agent-skills","ai-agents","browser-automation","claude-code","codex","html-to-markdown","playwright","web-automation","web-scraping"],"license":"mit","html_url":"https://github.com/appautomaton/webmaton","pushed_at":"2026-05-17T00:55:04Z","description":"Portable web-research and browser-automation SKILLs for Claude Code, Codex, and OpenCode — Playwright, Chrome DevTools, nodriver, and HTML-to-Markdown.","skill_md_sha":"40a8028603dc125bc39440c78f0917b56ba000d7","skill_md_path":"skills/agentic-search/SKILL.md","default_branch":"main","skill_tree_url":"https://github.com/appautomaton/webmaton/tree/main/skills/agentic-search"},"layout":"multi","source":"github","category":"webmaton","frontmatter":{"name":"agentic-search","description":"Grok-primary deep research skill for source-backed web work. Use when the task needs current web research, grounded citations, supplementary Tavily/Firecrawl source discovery, high-fidelity page-to-Markdown fetch, Tavily site mapping, verbatim quote extraction, source reranking, or reusable multi-step research sessions. Use as a supplement when ordinary search results feel incomplete, thinly sourced, or need more URLs, quotes, fetched pages, or session composition; not for simple one-off facts."},"skills_sh_url":"https://skills.sh/appautomaton/webmaton/agentic-search"},"updatedAt":"2026-05-18T19:06:02.127Z"}}