Internet research review

Internet Research Review

Use available web search or browser tools. Prefer direct public sources, cite links, and keep the output compact. Ask only when the topic, goal, time window, or locale materially changes the research.

Target

Infer or identify:

• Topic/entity: product, company, technology, error, market, person, competitor, feature, claim.
• Goal: news scan, community pain points, sentiment, issue triage, competitive intel, trend validation, source verification.
• Window: default last 30-90 days for current topics; expand only when history matters.
• Locale/language: use the user's locale when relevant; add English/Spanish variants for broader coverage.

Source Types

Choose only relevant categories. Do not hard-code sites unless the user names them or the topic has an obvious official source.

• Official documentation: product docs, API/SDK refs, integration/architecture guides, examples, FAQs, support KB.
• Official change/business: changelogs, release notes, version history, roadmap, status/incidents, deprecations, migrations, pricing, terms, privacy, security, compliance, filings, press rooms.
• Community: vendor/independent forums, Q&A communities, public discussion boards, technical help communities, user groups, public chat archives.
• Code/repositories: source repos, examples, templates, demos, manifests, commit history, PRs, discussions, wikis.
• Issues/support: bugs, feature requests, regressions, support tickets, labels, milestones, duplicates, closed/resolved issues, workarounds.
• Packages/ecosystem: registries, dependency pages, release metadata, download trends, maintainer notes, vulnerability notices.
• Tutorials/education: blog tutorials, courses, videos, workshops, conference talks, sample projects, implementation guides.
• News/analysis: industry/local media, newsletters, analyst notes, press releases, event coverage, launches, interviews.
• Reviews/marketplaces: software review portals, app stores, extension/plugin marketplaces, ecommerce reviews, ratings, comparisons.
• Social/professional: professional posts, short-form social, video comments, creator content, podcast notes, community recaps.
• Standards/governance: specs, RFCs, proposals, working-group notes, foundation pages, compatibility tables.
• Security/reliability: advisories, CVEs, exploit writeups, postmortems, uptime reports, dependency alerts, security policies.
• Competition/org signals: alternatives, comparison articles, migration posts, benchmarks, alternative directories, job posts, team pages, engineering blogs, public talks.

Search

Use grouped queries: "<topic>" + term.

• Official: documentation, API reference, changelog, release notes, status, migration guide.
• Community: forum, community, discussion, Q&A, user group, workaround.
• Code/issues: repository, issue, pull request, bug, regression.
• Tutorials: tutorial, example, guide, setup, best practices, implementation.
• Problems: error, complaints, broken, limitations, known issues.
• Adoption: launch, case study, benchmark, production, migration.
• Alternatives: alternative, vs, switching from, migrating from.
• Trust/risk: security advisory, vulnerability, outage, postmortem, pricing change.
• Spanish: documentacion, opiniones, problemas, foro, noticias, tutorial.

Use filters by page/source type when useful: documentation area, repo, issue tracker, discussion, forum, marketplace, news category, or official/vendor domain.

Quality

• Prefer primary/direct sources over snippets and aggregators.
• Cross-check key claims across at least two independent source types when possible.
• Separate official claims, user experience, speculation, and third-party analysis.
• Prefer recent active threads; treat old canonical docs differently from abandoned posts.
• Check closed/resolved issues to avoid reporting fixed problems as current.
• Check dates on tutorials; down-rank stale SEO pages unless evidence is specific.
• Down-rank copied press releases, anonymous speculation, AI summaries, and posts without first-hand evidence.
• Preserve credible contradictions and explain uncertainty.
• Do not access private, paywalled, credentialed, or sensitive content without explicit authorization.

Evidence

Capture: URL, title/source, author/org, date or "no visible date", claim/evidence, signal type, sentiment, confidence.

Signal types: official docs/change, community, tutorial, repository, issue, review, news, social, standard, security, competitor.

Rank by: primary evidence, repeated independent reports, specificity, impact, corroboration, freshness.

Output

Return a compact report:

• Executive signal: 2-4 strongest findings.
• Evidence table: source, date, signal, relevance, confidence, link.
• Patterns: repeated themes, disagreements, outliers.
• Risks: uncertainty, missing sources, bias, stale data.
• Next searches: only if more depth would materially improve confidence.

For monitoring requests, propose source categories, search terms, cadence, and thresholds. Do not create automations unless explicitly requested.