[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M78ASCTF70960ME151QZN7WV":3,"skill-same:01M78ASCTF70960ME151QZN7WV":84},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":17,"publishedAt":18,"linkOnly":19,"markdownContent":20,"markdownBody":21,"frontMatter":22,"refName":29,"commitSha":30,"skillDirectory":31,"forksCount":32,"repositoryHtmlUrl":33,"repositoryHomepageUrl":34,"ownerHtmlUrl":35,"ownerBlogUrl":6,"ownerDescription":36,"updatedAt":37,"install":38,"resources":51},"01M78ASCTF70960ME151QZN7WV","entity-research",null,"Use when the user says \"research this company \u002F person\", \"background check on X\", \"any negative press \u002F adverse media on X\", \"who owns \u002F who runs X\", \"is X sanctioned \u002F sanctions check on X\", \"vet this vendor \u002F counterparty \u002F candidate \u002F partner\", or \"entity search\" — deep background research on an entity (a company OR a person) into a cited dossier: identity & background, ownership & key management, adverse \u002F negative media, public sanctions-list name-match signals, PEP indications from public research, and litigation \u002F regulatory history. It RESEARCHES and COMPILES with sources; it is NOT a compliance\u002FAML\u002FCDD determination and NOT a sanctions or PEP clearance — a sanctions-list or PEP signal is a SIGNAL to escalate to a human compliance function, never a \"clear\" or \"block\". Research a person only for a legitimate purpose and only from public information.","moonlight-lupin\u002Fagent-skills","moonlight-lupin","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F136997773?v=4","01M78AEDSH1N62R0RNWPWMWVQY","01M78AEDT8ZF85MTK6XYBQBTNT",123,"MIT","https:\u002F\u002Fgithub.com\u002Fmoonlight-lupin\u002Fagent-skills\u002Fblob\u002Fmain\u002Fresearch\u002Fentity-research\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: entity-research\ndescription: \"Use when the user says \\\"research this company \u002F person\\\", \\\"background check on X\\\", \\\"any negative press \u002F adverse media on X\\\", \\\"who owns \u002F who runs X\\\", \\\"is X sanctioned \u002F sanctions check on X\\\", \\\"vet this vendor \u002F counterparty \u002F candidate \u002F partner\\\", or \\\"entity search\\\" — deep background research on an entity (a company OR a person) into a cited dossier: identity & background, ownership & key management, adverse \u002F negative media, public sanctions-list name-match signals, PEP indications from public research, and litigation \u002F regulatory history. It RESEARCHES and COMPILES with sources; it is NOT a compliance\u002FAML\u002FCDD determination and NOT a sanctions or PEP clearance — a sanctions-list or PEP signal is a SIGNAL to escalate to a human compliance function, never a \\\"clear\\\" or \\\"block\\\". Research a person only for a legitimate purpose and only from public information.\"\nlicense: MIT\nmetadata:\n  version: 1.2.0\n  author: moonlight-lupin\n  platforms: [linux, macos, windows]\n---\n\n# Entity Research\n\nDeep background research on a **company or a person** → a **cited dossier** for a human to read and act on. General-purpose research — vetting a vendor, a counterparty before a contract, a prospective hire, a partner, or verifying a media claim.\n\n> **Research & compilation — NOT a determination.** This skill does **not** screen, clear, rate, or block anyone. A sanctions-list \u002F PEP \u002F watchlist signal is a **SIGNAL to escalate** to a qualified compliance \u002F AML function, not a finding; negative press is an **allegation with a source and date**, not a proven fact. Everything is **cited**; nothing is auto-acted on.\n\n## Scope and routing\n\nUse this skill for general background research and fast public **signals**. Do **not** use it as a substitute for professional CDD\u002FAML screening, sanctions clearance, PEP screening, or a compliance decision, and do not build an intrusive profile of a private individual.\n\n## When to use it\n\n- Vet a **vendor \u002F supplier \u002F counterparty** before engaging or signing a contract.\n- Background on a **prospective hire**, a **partner**, or a co-investor's principal.\n- Check for **negative \u002F adverse media** or **litigation** on a name.\n- A quick **public sanctions-list signal** check (to escalate, not to clear).\n- A quick **PEP indication** check from public research (to escalate, not to clear).\n- \"Who actually owns \u002F runs X?\" — ownership & key-management background.\n\n## The research lenses\n\n1. **Identity & background** — confirm you have the right entity (registration no., jurisdiction, incorporation date, website, aliases \u002F former names; for a person: role, employer, location, DOB if public). Disambiguate same-name entities early.\n2. **Ownership & key management** — shareholders \u002F UBO signals, directors, senior managers; group\u002Fparent structure. Use the `people-enrichment` skill \u002F PDL for people & firmographics where appropriate.\n3. **Adverse \u002F negative media** — allegations, investigations, scandals, insolvency, fraud, environmental\u002Flabour issues — each with **source, date, and allegation-vs-outcome**.\n4. **Sanctions \u002F PEP \u002F watchlist signals** — `screen_lists(name)` checks public sanctions lists only (OFAC SDN + Consolidated, UK OFSI, UN). PEP indications are manual\u002Fopen-web research signals, such as public office, senior state-owned enterprise role, close-associate indications, or official biographies. **Neither is a clearance.**\n5. **Litigation & regulatory** — material lawsuits, regulator actions, fines, debarments.\n6. **Summary & flags** — a short read with **escalation flags** for a compliance reviewer.\n\n## Data sources (self-sufficient core + optional depth)\n\n- **Open web** — `web_search` + `web_extract` for press, litigation, registry mentions, ownership clues, and public PEP indications.\n- **Deep-research engine** — for a thorough pass, hand the entity + lenses to a deep-research engine\u002Fskill if available; otherwise run fan-out searches directly.\n- **PDL (people \u002F firmographics)** — run the **`people-enrichment`** skill for the owner \u002F key-management \u002F company layer where appropriate (needs `PDL_API_KEY`).\n- **Organisation evidence via the Exa Agent API (optional, paid)** — for the identity & background lens on **organisations only**: a hosted research agent returns the same cited fields (size, assets, mandate, sector exposure, recent commitments) for a batch of organisations, about US$0.10 a run (ceiling US$1.10; needs `EXA_API_KEY`). Never for people, never for screening, never with your relationship to the organisation in the task. Quote first; label and spot-check every fact; cite the underlying source. See `references\u002Forganisation-evidence-exa.md`.\n- **Public sanctions lists** — `scripts\u002Fentity_research.py` → `screen_lists(name)` fetches + token-name-matches the official government consolidated lists: **US (OFAC SDN + Consolidated), UK (OFSI), UN**. It returns **potential-match signals** only. It is not fuzzy\u002Fphonetic screening and does not cover all local autonomous lists. For a country's **local autonomous measures**, do a manual official-portal check. **A match is a signal to verify with a compliance function; no match is NOT a clearance**.\n\n## Workflow\n\n1. **Pin the subject** — name + identifiers (jurisdiction, registration no., website, role\u002Femployer for a person). Resolve same-name ambiguity before researching.\n2. **Plan the research** — break the entity into 3-6 research sub-questions across the lenses (e.g. \"Who owns X?\", \"Any litigation against X?\", \"Is X on any sanctions list?\"). Define success criteria: what would a complete dossier cover? This plan guides which lenses to prioritize and prevents skipping lenses.\n3. **Run the lenses** — fan-out web search per lens; PDL for people\u002Ffirmographics; `screen_lists()` for public sanctions-list signals; manual official\u002Fpublic checks for PEP indications and local sanctions lists. **Capture the source URL + date for every claim.**\n   - **Date grounding (mandatory).** Before searching, ground in the real current date: \"Today's date is {current date}. Use {current year} in queries — never a year inferred from training data.\"\n   - **Organisations with thin background facts** (or several needing the same facts) — offer the optional Exa evidence run with its cost, and run it only on a go-ahead (`references\u002Forganisation-evidence-exa.md`).\n   - **Quality filter.** Discard thin\u002Firrelevant results before extraction: landing pages, aggregator stubs (\u003C100 words of substantive content), pages with keyword overlap but no actual relevance (word-boundary match entity name, not substring), and duplicate URLs.\n4. **Gap analysis** — after the first pass, review findings against the research plan from step 2. Which sub-questions are unanswered? Which lenses have thin coverage? Generate targeted follow-up queries for the gaps and run a second search pass. Repeat once more if significant gaps remain (max 3 passes).\n5. **Weigh** — primary\u002Fofficial sources > reputable press > blogs\u002Fforums; allegation vs outcome; recency; corroboration (≥2 independent sources for a serious claim). Flag low-confidence items as such; don't launder rumour into fact. **Little or no public trace:** when the subject barely exists online (no registry hit, no website, no press), list every check that came back empty — which registry, which list, which search, on what date — and state plainly what cannot be verified. Silence is never evidence that the entity exists, is legitimate or is clean; ask the user for documents (certificate of incorporation, register of directors) and flag it for review.\n6. **Assemble** — `dossier(...)` builds the cited markdown dossier (the six lenses + a \"not a determination\" header + escalation flags). Deliver to chat or save as `.md`.\n7. **Flag escalations** — any sanctions-list signal, PEP indication, serious adverse finding, or integrity concern → call out **\"escalate to a compliance \u002F AML reviewer\"** explicitly.\n\nSee `references\u002Fresearch-lenses.md` (what to look for + query patterns) and `references\u002Fboundaries-and-sanctions.md` (the signal-not-determination rule, false positives, and privacy guardrails).\n\n## Output pattern\n\nFor adverse \u002F litigation \u002F regulatory findings, prefer a compact table:\n\n| Item | Source | Date | Allegation \u002F charge \u002F outcome | Confidence | Escalate? |\n| --- | --- | --- | --- | --- | --- |\n\nFor sanctions \u002F PEP \u002F watchlist signals, never write \"clear\" or \"blocked\". Use language like:\n\n- `Public sanctions-list check found no potential matches in OFAC SDN\u002FConsolidated, UK OFSI and UN via screen_lists() as of [date]. This is not a clearance; local lists and professional screening remain outside this helper.`\n- `Potential sanctions-list name match on \"[matched name]\" ([list], score [x]) — escalate to compliance \u002F AML reviewer for verification.`\n- `Public PEP indication: [public office \u002F SOE role \u002F close associate indication] from [source, date] — escalate for compliance review. This is not a PEP-screening determination.`\n\n## Boundaries & safety\n\n- **Research, not a determination \u002F clearance.** Sanctions-list or PEP signal = **escalate**, never \"clear\" or \"block\". No match ≠ clean.\n- **PEP clarity.** `screen_lists()` is not a PEP screener. PEP indications come from public-source research and must be reviewed by a qualified compliance function.\n- **Allegations vs facts.** Attribute and date every negative item; distinguish allegation, charge, and outcome. Avoid defamatory framing; report what sources say, with the source.\n- **Persons — legitimate purpose, public info only.** Research a person only for a legitimate purpose (vetting), and only **publicly-available** information; don't compile sensitive personal data (health, beliefs, sexuality, etc.) or build an intrusive profile.\n- **A dossier is a draft for a human** — never the basis for an automated action.\n\n## Principles\n\n- **Drafts, not advice** — a dossier is a research aid for a person to read and act on.\n- **Never invent** — cite every claim with a source + date; mark thin\u002Funcorroborated items as such.\n- **Signal, not determination** — a sanctions-list \u002F PEP \u002F watchlist signal escalates; it never clears or blocks.\n- **Honesty and calibration** — distinguish allegation from outcome; present conflicts, note confidence.\n- **Workspace hygiene** — keep the dossier local; it's internal and may name individuals.\n\n## Data handling — search the name, not the relationship\n\nA **bare name with no relationship attached is fine** to research on the open web. But:\n\n- **Never leak the context.** Search \"[entity]\", not \"we're investing in [entity]\" or \"[entity] our client\" — keep your deal\u002Fclient relationship out of external queries.\n- If the entity **is** tied to a live deal or a client, the relationship stays confidential (omit it from queries); the public research on the name still proceeds.\n- Keep the dossier on the local machine; it may name individuals.\n\n## Files\n\n- `scripts\u002Fentity_research.py` — `screen_lists` (public sanctions-list name-match signal, cached, graceful), `dossier` (assemble the cited markdown dossier + escalation flags), name-normalisation\u002Fcitation helpers; `--self-test` (offline; matcher + dossier).\n- `references\u002Fresearch-lenses.md` — per-lens checklist + good query patterns + source weighting.\n- `references\u002Fboundaries-and-sanctions.md` — the lists, signal-not-determination, false positives, escalation, and person-privacy guardrails.\n- `references\u002Forganisation-evidence-exa.md` — the optional Exa Agent API run for cited organisation facts: when to use it, the US$1 minimum budget cap, task wording, labelling and spot-checks.\n\n## Verification checklist\n\n- [ ] Subject pinned; same-name ambiguity resolved (or both presented).\n- [ ] Research plan created (3-6 sub-questions across lenses, success criteria defined).\n- [ ] Date grounding applied — queries use current year, not training-cutoff year.\n- [ ] Quality filter applied — thin\u002Firrelevant\u002Fduplicate results discarded before extraction.\n- [ ] Gap analysis run — at least 2 search passes; remaining gaps documented.\n- [ ] Every claim carries a source URL + date; serious claims corroborated (≥2 sources).\n- [ ] Allegations attributed and distinguished from outcomes; thin items flagged as thin.\n- [ ] `screen_lists()` presented as public sanctions-list signal only, never PEP screening or clearance.\n- [ ] PEP indications, if any, came from public sources and are presented as escalation signals only.\n- [ ] Person research limited to a legitimate purpose and public info only.\n- [ ] Escalation flags listed explicitly; dossier kept local.\n- [ ] If the Exa evidence run was used: organisations only, cost quoted and approved, `usage.emails`\u002F`phoneNumbers` 0, each fact labelled and at least one spot-checked, the underlying source cited.\n\n## Requirements\n\n- Python 3.8+ (stdlib only for `screen_lists`\u002F`dossier`).\n- Session **web search + fetch** tools for the research lenses (not bundled).\n- **Network** for `screen_lists` (public lists) — `--self-test` runs offline.\n- Optional: `PDL_API_KEY` + the `people-enrichment` skill for the people\u002Ffirmographics layer; a deep-research engine\u002Fskill for a deeper pass; `EXA_API_KEY` for the organisation-evidence run.\n","# Entity Research\n\nDeep background research on a **company or a person** → a **cited dossier** for a human to read and act on. General-purpose research — vetting a vendor, a counterparty before a contract, a prospective hire, a partner, or verifying a media claim.\n\n> **Research & compilation — NOT a determination.** This skill does **not** screen, clear, rate, or block anyone. A sanctions-list \u002F PEP \u002F watchlist signal is a **SIGNAL to escalate** to a qualified compliance \u002F AML function, not a finding; negative press is an **allegation with a source and date**, not a proven fact. Everything is **cited**; nothing is auto-acted on.\n\n## Scope and routing\n\nUse this skill for general background research and fast public **signals**. Do **not** use it as a substitute for professional CDD\u002FAML screening, sanctions clearance, PEP screening, or a compliance decision, and do not build an intrusive profile of a private individual.\n\n## When to use it\n\n- Vet a **vendor \u002F supplier \u002F counterparty** before engaging or signing a contract.\n- Background on a **prospective hire**, a **partner**, or a co-investor's principal.\n- Check for **negative \u002F adverse media** or **litigation** on a name.\n- A quick **public sanctions-list signal** check (to escalate, not to clear).\n- A quick **PEP indication** check from public research (to escalate, not to clear).\n- \"Who actually owns \u002F runs X?\" — ownership & key-management background.\n\n## The research lenses\n\n1. **Identity & background** — confirm you have the right entity (registration no., jurisdiction, incorporation date, website, aliases \u002F former names; for a person: role, employer, location, DOB if public). Disambiguate same-name entities early.\n2. **Ownership & key management** — shareholders \u002F UBO signals, directors, senior managers; group\u002Fparent structure. Use the `people-enrichment` skill \u002F PDL for people & firmographics where appropriate.\n3. **Adverse \u002F negative media** — allegations, investigations, scandals, insolvency, fraud, environmental\u002Flabour issues — each with **source, date, and allegation-vs-outcome**.\n4. **Sanctions \u002F PEP \u002F watchlist signals** — `screen_lists(name)` checks public sanctions lists only (OFAC SDN + Consolidated, UK OFSI, UN). PEP indications are manual\u002Fopen-web research signals, such as public office, senior state-owned enterprise role, close-associate indications, or official biographies. **Neither is a clearance.**\n5. **Litigation & regulatory** — material lawsuits, regulator actions, fines, debarments.\n6. **Summary & flags** — a short read with **escalation flags** for a compliance reviewer.\n\n## Data sources (self-sufficient core + optional depth)\n\n- **Open web** — `web_search` + `web_extract` for press, litigation, registry mentions, ownership clues, and public PEP indications.\n- **Deep-research engine** — for a thorough pass, hand the entity + lenses to a deep-research engine\u002Fskill if available; otherwise run fan-out searches directly.\n- **PDL (people \u002F firmographics)** — run the **`people-enrichment`** skill for the owner \u002F key-management \u002F company layer where appropriate (needs `PDL_API_KEY`).\n- **Organisation evidence via the Exa Agent API (optional, paid)** — for the identity & background lens on **organisations only**: a hosted research agent returns the same cited fields (size, assets, mandate, sector exposure, recent commitments) for a batch of organisations, about US$0.10 a run (ceiling US$1.10; needs `EXA_API_KEY`). Never for people, never for screening, never with your relationship to the organisation in the task. Quote first; label and spot-check every fact; cite the underlying source. See `references\u002Forganisation-evidence-exa.md`.\n- **Public sanctions lists** — `scripts\u002Fentity_research.py` → `screen_lists(name)` fetches + token-name-matches the official government consolidated lists: **US (OFAC SDN + Consolidated), UK (OFSI), UN**. It returns **potential-match signals** only. It is not fuzzy\u002Fphonetic screening and does not cover all local autonomous lists. For a country's **local autonomous measures**, do a manual official-portal check. **A match is a signal to verify with a compliance function; no match is NOT a clearance**.\n\n## Workflow\n\n1. **Pin the subject** — name + identifiers (jurisdiction, registration no., website, role\u002Femployer for a person). Resolve same-name ambiguity before researching.\n2. **Plan the research** — break the entity into 3-6 research sub-questions across the lenses (e.g. \"Who owns X?\", \"Any litigation against X?\", \"Is X on any sanctions list?\"). Define success criteria: what would a complete dossier cover? This plan guides which lenses to prioritize and prevents skipping lenses.\n3. **Run the lenses** — fan-out web search per lens; PDL for people\u002Ffirmographics; `screen_lists()` for public sanctions-list signals; manual official\u002Fpublic checks for PEP indications and local sanctions lists. **Capture the source URL + date for every claim.**\n   - **Date grounding (mandatory).** Before searching, ground in the real current date: \"Today's date is {current date}. Use {current year} in queries — never a year inferred from training data.\"\n   - **Organisations with thin background facts** (or several needing the same facts) — offer the optional Exa evidence run with its cost, and run it only on a go-ahead (`references\u002Forganisation-evidence-exa.md`).\n   - **Quality filter.** Discard thin\u002Firrelevant results before extraction: landing pages, aggregator stubs (\u003C100 words of substantive content), pages with keyword overlap but no actual relevance (word-boundary match entity name, not substring), and duplicate URLs.\n4. **Gap analysis** — after the first pass, review findings against the research plan from step 2. Which sub-questions are unanswered? Which lenses have thin coverage? Generate targeted follow-up queries for the gaps and run a second search pass. Repeat once more if significant gaps remain (max 3 passes).\n5. **Weigh** — primary\u002Fofficial sources > reputable press > blogs\u002Fforums; allegation vs outcome; recency; corroboration (≥2 independent sources for a serious claim). Flag low-confidence items as such; don't launder rumour into fact. **Little or no public trace:** when the subject barely exists online (no registry hit, no website, no press), list every check that came back empty — which registry, which list, which search, on what date — and state plainly what cannot be verified. Silence is never evidence that the entity exists, is legitimate or is clean; ask the user for documents (certificate of incorporation, register of directors) and flag it for review.\n6. **Assemble** — `dossier(...)` builds the cited markdown dossier (the six lenses + a \"not a determination\" header + escalation flags). Deliver to chat or save as `.md`.\n7. **Flag escalations** — any sanctions-list signal, PEP indication, serious adverse finding, or integrity concern → call out **\"escalate to a compliance \u002F AML reviewer\"** explicitly.\n\nSee `references\u002Fresearch-lenses.md` (what to look for + query patterns) and `references\u002Fboundaries-and-sanctions.md` (the signal-not-determination rule, false positives, and privacy guardrails).\n\n## Output pattern\n\nFor adverse \u002F litigation \u002F regulatory findings, prefer a compact table:\n\n| Item | Source | Date | Allegation \u002F charge \u002F outcome | Confidence | Escalate? |\n| --- | --- | --- | --- | --- | --- |\n\nFor sanctions \u002F PEP \u002F watchlist signals, never write \"clear\" or \"blocked\". Use language like:\n\n- `Public sanctions-list check found no potential matches in OFAC SDN\u002FConsolidated, UK OFSI and UN via screen_lists() as of [date]. This is not a clearance; local lists and professional screening remain outside this helper.`\n- `Potential sanctions-list name match on \"[matched name]\" ([list], score [x]) — escalate to compliance \u002F AML reviewer for verification.`\n- `Public PEP indication: [public office \u002F SOE role \u002F close associate indication] from [source, date] — escalate for compliance review. This is not a PEP-screening determination.`\n\n## Boundaries & safety\n\n- **Research, not a determination \u002F clearance.** Sanctions-list or PEP signal = **escalate**, never \"clear\" or \"block\". No match ≠ clean.\n- **PEP clarity.** `screen_lists()` is not a PEP screener. PEP indications come from public-source research and must be reviewed by a qualified compliance function.\n- **Allegations vs facts.** Attribute and date every negative item; distinguish allegation, charge, and outcome. Avoid defamatory framing; report what sources say, with the source.\n- **Persons — legitimate purpose, public info only.** Research a person only for a legitimate purpose (vetting), and only **publicly-available** information; don't compile sensitive personal data (health, beliefs, sexuality, etc.) or build an intrusive profile.\n- **A dossier is a draft for a human** — never the basis for an automated action.\n\n## Principles\n\n- **Drafts, not advice** — a dossier is a research aid for a person to read and act on.\n- **Never invent** — cite every claim with a source + date; mark thin\u002Funcorroborated items as such.\n- **Signal, not determination** — a sanctions-list \u002F PEP \u002F watchlist signal escalates; it never clears or blocks.\n- **Honesty and calibration** — distinguish allegation from outcome; present conflicts, note confidence.\n- **Workspace hygiene** — keep the dossier local; it's internal and may name individuals.\n\n## Data handling — search the name, not the relationship\n\nA **bare name with no relationship attached is fine** to research on the open web. But:\n\n- **Never leak the context.** Search \"[entity]\", not \"we're investing in [entity]\" or \"[entity] our client\" — keep your deal\u002Fclient relationship out of external queries.\n- If the entity **is** tied to a live deal or a client, the relationship stays confidential (omit it from queries); the public research on the name still proceeds.\n- Keep the dossier on the local machine; it may name individuals.\n\n## Files\n\n- `scripts\u002Fentity_research.py` — `screen_lists` (public sanctions-list name-match signal, cached, graceful), `dossier` (assemble the cited markdown dossier + escalation flags), name-normalisation\u002Fcitation helpers; `--self-test` (offline; matcher + dossier).\n- `references\u002Fresearch-lenses.md` — per-lens checklist + good query patterns + source weighting.\n- `references\u002Fboundaries-and-sanctions.md` — the lists, signal-not-determination, false positives, escalation, and person-privacy guardrails.\n- `references\u002Forganisation-evidence-exa.md` — the optional Exa Agent API run for cited organisation facts: when to use it, the US$1 minimum budget cap, task wording, labelling and spot-checks.\n\n## Verification checklist\n\n- [ ] Subject pinned; same-name ambiguity resolved (or both presented).\n- [ ] Research plan created (3-6 sub-questions across lenses, success criteria defined).\n- [ ] Date grounding applied — queries use current year, not training-cutoff year.\n- [ ] Quality filter applied — thin\u002Firrelevant\u002Fduplicate results discarded before extraction.\n- [ ] Gap analysis run — at least 2 search passes; remaining gaps documented.\n- [ ] Every claim carries a source URL + date; serious claims corroborated (≥2 sources).\n- [ ] Allegations attributed and distinguished from outcomes; thin items flagged as thin.\n- [ ] `screen_lists()` presented as public sanctions-list signal only, never PEP screening or clearance.\n- [ ] PEP indications, if any, came from public sources and are presented as escalation signals only.\n- [ ] Person research limited to a legitimate purpose and public info only.\n- [ ] Escalation flags listed explicitly; dossier kept local.\n- [ ] If the Exa evidence run was used: organisations only, cost quoted and approved, `usage.emails`\u002F`phoneNumbers` 0, each fact labelled and at least one spot-checked, the underlying source cited.\n\n## Requirements\n\n- Python 3.8+ (stdlib only for `screen_lists`\u002F`dossier`).\n- Session **web search + fetch** tools for the research lenses (not bundled).\n- **Network** for `screen_lists` (public lists) — `--self-test` runs offline.\n- Optional: `PDL_API_KEY` + the `people-enrichment` skill for the people\u002Ffirmographics layer; a deep-research engine\u002Fskill for a deeper pass; `EXA_API_KEY` for the organisation-evidence run.\n",{"name":5,"license":14,"metadata":23,"description":7},{"author":9,"version":24,"platforms":25},"1.2.0",[26,27,28],"linux","macos","windows","main","e943e3285cd03c44b2c762a9a52328e3d16444b1","research\u002Fentity-research",22,"https:\u002F\u002Fgithub.com\u002Fmoonlight-lupin\u002Fagent-skills","https:\u002F\u002Fhermes-agent.nousresearch.com","https:\u002F\u002Fgithub.com\u002Fmoonlight-lupin","Accounting\u002FFP&A\u002FM&A\u002FPERE guy by day, AI\u002Fself-host\u002Fprogramming hobbyist by night. \r\n\r\nPassionate in solving practical business issues.","2026-11-13 19:46:18",{"source":8,"skillName":5,"commands":39,"prompt":49,"zipAvailable":50},[40,43,46],{"tool":41,"command":42},"npx","npx skills add moonlight-lupin\u002Fagent-skills --skill entity-research",{"tool":44,"command":45},"bunx","bunx skills add moonlight-lupin\u002Fagent-skills --skill entity-research",{"tool":47,"command":48},"pnpm","pnpm dlx skills add moonlight-lupin\u002Fagent-skills --skill entity-research","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fmoonlight-lupin\u002Fagent-skills\u002Fblob\u002Fmain\u002Fresearch\u002Fentity-research\u002FSKILL.md",true,[52,58,62,66,70,76,80],{"path":53,"resourceType":54,"resourceTypeDesc":55,"sizeBytes":56,"sourceHtmlUrl":57},"README.md","REFERENCE","参考文档",1809,"https:\u002F\u002Fgithub.com\u002Fmoonlight-lupin\u002Fagent-skills\u002Fblob\u002Fmain\u002Fresearch\u002Fentity-research\u002FREADME.md",{"path":59,"resourceType":54,"resourceTypeDesc":55,"sizeBytes":60,"sourceHtmlUrl":61},"references\u002Fboundaries-and-sanctions.md",4599,"https:\u002F\u002Fgithub.com\u002Fmoonlight-lupin\u002Fagent-skills\u002Fblob\u002Fmain\u002Fresearch\u002Fentity-research\u002Freferences\u002Fboundaries-and-sanctions.md",{"path":63,"resourceType":54,"resourceTypeDesc":55,"sizeBytes":64,"sourceHtmlUrl":65},"references\u002Forganisation-evidence-exa.md",5828,"https:\u002F\u002Fgithub.com\u002Fmoonlight-lupin\u002Fagent-skills\u002Fblob\u002Fmain\u002Fresearch\u002Fentity-research\u002Freferences\u002Forganisation-evidence-exa.md",{"path":67,"resourceType":54,"resourceTypeDesc":55,"sizeBytes":68,"sourceHtmlUrl":69},"references\u002Fresearch-lenses.md",3795,"https:\u002F\u002Fgithub.com\u002Fmoonlight-lupin\u002Fagent-skills\u002Fblob\u002Fmain\u002Fresearch\u002Fentity-research\u002Freferences\u002Fresearch-lenses.md",{"path":71,"resourceType":72,"resourceTypeDesc":73,"sizeBytes":74,"sourceHtmlUrl":75},"scripts\u002Fentity_research.py","SCRIPT","脚本",17523,"https:\u002F\u002Fgithub.com\u002Fmoonlight-lupin\u002Fagent-skills\u002Fblob\u002Fmain\u002Fresearch\u002Fentity-research\u002Fscripts\u002Fentity_research.py",{"path":77,"resourceType":72,"resourceTypeDesc":73,"sizeBytes":78,"sourceHtmlUrl":79},"tests\u002F__init__.py",0,"https:\u002F\u002Fgithub.com\u002Fmoonlight-lupin\u002Fagent-skills\u002Fblob\u002Fmain\u002Fresearch\u002Fentity-research\u002Ftests\u002F__init__.py",{"path":81,"resourceType":72,"resourceTypeDesc":73,"sizeBytes":82,"sourceHtmlUrl":83},"tests\u002Ftest_entity_research.py",23972,"https:\u002F\u002Fgithub.com\u002Fmoonlight-lupin\u002Fagent-skills\u002Fblob\u002Fmain\u002Fresearch\u002Fentity-research\u002Ftests\u002Ftest_entity_research.py",[85,92,99,106,113,120],{"skillNo":86,"name":87,"nameZh":6,"description":88,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":89,"categories":90,"keywords":91,"publishedAt":18},"01M78ATNRV7RYQ8M314V0QZX18","website-scraping","Generic playbook for extracting structured data from any website — hotel prices, flight fares, e-commerce listings, real estate, jobs, competitor product catalogues, anything where the goal is to turn one or more URLs into clean records on disk. Use this skill whenever the user mentions scraping, harvesting, extracting, or pulling data from a website; names a specific site or competitor they want data from; asks how to handle JavaScript-rendered pages, Cloudflare blocks, bot detection, Turnstile challenges, or Playwright; wants to monitor prices over time; needs to automate a copy-paste research task; or says things like \"just get the data from X\" or \"I need a script that grabs N from site Y\". Covers recon, picking the lightest extraction tool that works, surviving anti-bot defences, and writing clean JSONL output with a run manifest. Scope is scraping only — landing data in a database, scheduling, and downstream pipeline work are explicitly out of scope and live in other tools.","https:\u002F\u002Fgithub.com\u002Fmoonlight-lupin\u002Fagent-skills\u002Fblob\u002Fmain\u002Fweb-scraping\u002Fwebsite-scraping\u002FSKILL.md",[],[],{"skillNo":93,"name":94,"nameZh":6,"description":95,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":96,"categories":97,"keywords":98,"publishedAt":18},"01M78ATK9XGEKSH3QRN7P2MEER","greet","Greets a named person warmly.","https:\u002F\u002Fgithub.com\u002Fmoonlight-lupin\u002Fagent-skills\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Fclaude_plugin_sample\u002Fskills\u002Fgreet\u002FSKILL.md",[],[],{"skillNo":100,"name":101,"nameZh":6,"description":102,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":103,"categories":104,"keywords":105,"publishedAt":18},"01M78ATEAQB5CK2D4MKPCW4YHQ","skill-quality-review","Use when auditing or reviewing skill quality — batch checks across a library, deep per-skill grep methodology, usage stats, efficacy testing, or fix planning. Single consolidated home for the writing-for-agents review framework and the 3-surface review methodology.","https:\u002F\u002Fgithub.com\u002Fmoonlight-lupin\u002Fagent-skills\u002Fblob\u002Fmain\u002Fsoftware-development\u002Fskill-quality-review\u002FSKILL.md",[],[],{"skillNo":107,"name":108,"nameZh":6,"description":109,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":110,"categories":111,"keywords":112,"publishedAt":18},"01M78AT58EM57023QSC79S8R6D","people-enrichment","Use when the user has a list of names or companies and wants their job, employer, work history, LinkedIn URLs or company profiles looked up, or wants to find people or companies matching criteria — trigger on phrases like \"enrich these contacts\", \"find their LinkedIn\", \"look up these people\", \"who works at X\", \"find directors at Y\", \"enrich these companies\", or \"get firmographics for\". Enrich and search People Data Labs (PDL) Person and Company data, writing results to .xlsx. Five operations: enrich named PEOPLE into a profile, employment history and LinkedIn URL; IDENTIFY an ambiguous person as several scored candidates; SEARCH people by criteria (company, title, location); enrich COMPANIES into firmographics (industry, size, employees, HQ, founded, LinkedIn); and SEARCH companies by criteria. Do not scrape LinkedIn directly — this uses a licensed data aggregator (PDL) instead, for legal and reliability reasons.","https:\u002F\u002Fgithub.com\u002Fmoonlight-lupin\u002Fagent-skills\u002Fblob\u002Fmain\u002Fresearch\u002Fpeople-enrichment\u002FSKILL.md",[],[],{"skillNo":114,"name":115,"nameZh":6,"description":116,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":117,"categories":118,"keywords":119,"publishedAt":18},"01M78AT0AZJ7WFY73766NBA1EC","notebooklm-mode","Source-grounded research pipeline inspired by NotebookLM. Collect verbatim source extracts into a temporary workspace, then answer questions and create deliverables grounded in those sources. Two grounding modes: strict (vault-only) and augmented (vault + labeled [background] knowledge). Trigger when the user asks to research a topic from sources, wants NotebookLM-style grounded answers, asks to build a resource vault, or says \"notebooklm mode\".","https:\u002F\u002Fgithub.com\u002Fmoonlight-lupin\u002Fagent-skills\u002Fblob\u002Fmain\u002Fresearch\u002Fnotebooklm-mode\u002FSKILL.md",[],[],{"skillNo":121,"name":122,"nameZh":6,"description":123,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":124,"categories":125,"keywords":126,"publishedAt":18},"01M78ASTJVZ9G5JY6GM40SF909","news-monitoring","Use when a topic or news stream needs recurring monitoring — cover it with web search, multi-language sources, digest formatting, and automated delivery via Hermes cron jobs. Covers search strategy, source selection, Chinese-language platforms, and digest templates.","https:\u002F\u002Fgithub.com\u002Fmoonlight-lupin\u002Fagent-skills\u002Fblob\u002Fmain\u002Fresearch\u002Fnews-monitoring\u002FSKILL.md",[],[]]