octocode-scraping
Use when extracting or mapping public web content into a local cited corpus: scrape or crawl a URL/docs site, pull tables/pricing/product fields, diagnose blocked or thin pages, or answer from saved pages. Phrases like scrape this URL, crawl the docs, build a corpus, extract pricing. Prefer keyless fetch; ask before hosted spend. Live clicks/HAR/perf → octocode-chrome-devtools.
How do I install this agent skill?
npx skills add https://github.com/bgauryy/octocode --skill octocode-scrapingIs this agent skill safe to install?
- Gen Agent Trust Hubpass
The skill is a comprehensive web scraping and crawling utility. It emphasizes security and privacy through automated redaction of sensitive data (cookies, auth headers, tokens) and includes explicit policies for handling untrusted content. While it possesses powerful capabilities like local script execution and broad web ingestion, these are architected with developer-focused controls and safety warnings.
- Socketwarn
2 alerts: gptAnomaly
- Snykwarn
Risk: MEDIUM · 1 issue
What does this agent skill do?
Octocode Scraping
Flow: FRAME → POLICY → ROUTE → FETCH → CORPUS → SEARCH → CITE → RECOVER.
FRAME before the first fetch: fix target URL/domain, goal, depth, and output shape — vague ask → references/user-inputs.md.
Defaults: one public URL, --mode html, omit --provider (keyless cdp→direct), session .octocode/tmp/scrape/{sessionId}, compact stdout. Search corpus before refetch. Live interaction → chrome-devtools on one port, then har-ingest + corpus-run into the same session. Ask before auth, hosted spend, crawl widen, CAPTCHA/MFA, destructive actions. Cite paths + URL metadata — not raw dumps.
Stop when: two same-class failures (report evidence, route tried, sanitized status, next approval); hosted 403 (wrong key or credits gone — status only, no retry); CAPTCHA/MFA, auth wall, or cookie/profile transfer needed; still blocked after one cdp try (ask before --provider scrapingant); personal data, form submits, purchases, sends, deletes, or account changes in scope; the saved corpus already proves the claim (cite it, do not refetch); crawl widen before reports/summary.md is useful. Recovery table: references/failure-recovery.md.
Route (pick one)
| Need | Do | Skip |
|---|---|---|
| Vague scrape | --mode html, omit --provider | markdown / auto hosted |
| See auto pick | scripts/provider-check.mjs | guessing |
| Fetch/crawl | scripts/fetch.mjs | deprecated scrapingant-fetch |
| Corpus on disk | corpus-inspect → find helpers → corpus-run | blind refetch |
| Live click/DOM/auth | chrome-devtools | --provider cdp alone for interaction |
| Page health | chrome measure + measure-query | hosted scrape for scores |
| CDP → corpus | scripts/har-ingest.mjs --from-cdp-dir | new sessionId |
| Prove field | scripts/corpus-run.mjs --regex|--script | reopen browser |
| Still blocked | evidence + ask → --provider scrapingant | silent spend |
Scripts (Node, no install; every one takes --help)
| When | Run |
|---|---|
| fetch / crawl / extract a URL — the owner of every network call | scripts/fetch.mjs --url <u> [--mode html] [--crawl --same-domain --max-pages <n>] [--no-raw] |
| want the fetch plus an immediate corpus brief in one shot | scripts/fetch-and-brief.mjs --url <u> (wraps fetch.mjs → corpus-inspect.mjs) |
| before routing or hosted spend: which provider auto-wins, credits left | scripts/provider-check.mjs [--provider <p>], scripts/provider-usage.mjs — sanitized, never prints the key |
| read a saved session first, then search it | scripts/corpus-inspect.mjs --session-dir <d> [--page <n>], then scripts/corpus-find.mjs --session-dir <d> --query <t> |
| pull static DOM, assets, or graph paths out of the corpus (live DOM → chrome-devtools) | scripts/dom-find.mjs --kind form/button/table, scripts/resource-list.mjs --kind asset/external, scripts/graph-navigate.mjs --from <nodeId> — each with --session-dir <d> |
| prove one field locally instead of reopening a browser | scripts/corpus-run.mjs --session-dir <d> --roots cdp,extracts --regex <re> (or --script <file>, --concat-parts) |
| bridge chrome-devtools artifacts into this session, or hand a packet back | scripts/har-ingest.mjs --session-dir <d> --from-cdp-dir <run>; reverse with --export-packet |
| need field names before an extraction | scripts/schema-helper.mjs --intent "extract pricing and features" |
| an old transcript uses the legacy names | deprecated shims scripts/scrapingant-fetch.mjs, scripts/scrapingant-check.mjs, scripts/scrapingant-usage.mjs forward to fetch.mjs / provider-check.mjs / provider-usage.mjs — call the new names |
| editing or extending a script | shared modules in scripts/lib/ (providers registry, client fetch, corpus, analyzers, extractors, text, args, bridge); env/key resolution vendored in scripts/octocode-config.mjs; JSON contracts in scripts/schemas/ |
Full flags, roles, and the library map: read scripts/README.md before adding a flag or a vendor.
References
| When | Load |
|---|---|
| frame a vague scope | references/user-inputs.md |
| legal/safety/privacy policy | references/scraping-policy.md |
| cost / keyless vs hosted route | references/route-selection.md |
| provider registry / add a vendor | references/providers.md |
| hosted API, after approval | references/scrapingant.md |
| search a saved session corpus | references/session-corpus.md |
| site graph / navigation / workflows | references/website-analysis.md |
| stdout and corpus file shapes | references/data-contract.md |
| extract quality and citations | references/extraction-quality.md |
| live browser bridge playbook | references/browser-scraping.md |
| blocked, thin, or oversized results | references/failure-recovery.md |
How can the creator link this skill?
Add the canonical catalog link to the repository README so users can inspect current installs and available audits. The publishing guide covers the complete discovery path.
<a href="https://skillzs.dev/skills/bgauryy/octocode/octocode-scraping">View octocode-scraping on skillZs</a>