people-search
Find and rank professional people for recruiting, partnerships, sales, or research from user-provided data, public web sources, an authenticated search session, or a connected provider. Use for people discovery, LinkedIn or Sales Navigator search design, profile-list ranking, or provider filter translation.
- License 许可
- Apache-2.0
- 所属仓库
- getedgehq/skills
- 最近更新
- 2026-11-12
- Stars
- ★ 5
- Forks
- 0
01 前言 Front Matter
- name
- people-search
- description
- Find and rank professional people for recruiting, partnerships, sales, or research from user-provided data, public web sources, an authenticated search session, or a connected provider. Use for people discovery, LinkedIn or Sales Navigator search design, profile-list ranking, or provider filter translation.
02 资源文件 8
- 参考文档agents/openai.yaml205 B
- 参考文档DERIVATION.json4.7 KB
- 许可证LICENSE10.5 KB
- 参考文档references/provider-contract.md3.6 KB
- 参考文档references/search-schema.md1.1 KB
- 脚本scripts/people_search.py22.1 KB
- 资产文件tests/fixtures/people.csv632 B
- 脚本tests/test_people_search.py15.9 KB
03 SKILL.md 正文
People Search
Produce an auditable shortlist without pretending the skill itself grants LinkedIn data access.
Choose the data mode
Use the strongest mode already available: provided CSV/profiles/export; public web search; an authorized authenticated session; or a connected provider such as HarvestAPI or Apify. If no discovery source is available, return a normalized search plan, ready-to-run filters, and an input template. Never fabricate profiles.
Workflow
- Translate the request into the canonical schema.
- Separate hard requirements from ranking preferences. Treat vague criteria as preferences unless explicitly mandatory.
- State the selected mode and its coverage limits.
- Map canonical filters to provider fields and disclose every applied, approximated, dropped, or unsupported filter.
- Run the cheapest useful canary before a paid/full search. Stop on invalid, silently empty, or off-target results.
- Deduplicate by stable profile URL, then normalized name plus current company.
- Rank against the original brief, preserve evidence URLs, and distinguish sourced facts from inference.
- Return the shortlist, search recipe, caveats, and highest-value refinement.
Boundaries
- A skill is orchestration, not a data license. Never imply built-in LinkedIn access.
- Never expose, embed, or persist provider credentials.
- Make the cost-bearing mode clear before paid search; prefer previews and staged expansion.
- Never interpret zero results as no matching people until request validity and filter support are checked.
- Do not send connection requests or messages without separate authorization.
For adapter behavior and result contracts, read references/provider-contract.md.
Deterministic runner
Use scripts/people_search.py for plan-only, CSV import/ranking, and provider execution. Provider credentials must enter through the named environment variable at process scope; never pass a token as an argument.
Hard filters are field-scoped: --locations matches the location field only, --titles the current_title, --companies the current_company, --industries the industry. Only --must-keywords and --exclude-keywords scan several fields, and those are reported as approximated. Exclusions are first-class: --exclude-companies, --exclude-titles, --exclude-keywords, --exclude-profile-urls.
Every requested filter appears exactly once in filters, sorted into applied, approximated, dropped, or unsupported, with the fields it matched and the reason for its status. A filter whose target field is missing from the data is dropped and not enforced, so the result set is wider than the brief and says so. Report that ledger to the user; do not restate the brief as if all of it ran.
Provider runs are staged. --provider-run canary is the default: a preview request capped by --canary-limit (5), never a complete result set. --provider-run full sends preview: false and the real --limit, and requires --max-cost-usd. Add --cost-per-result-usd so the cap is checked before the request instead of after the invoice; without it cost.cap_enforced is false. A provider-reported cost above the cap comes back in warnings.
Read ok before results. On success ok is true, results_valid is true, and results is a list that may legitimately be empty, in which case zero_result_note and diagnostics.rejected_by_filter explain what removed the candidates. On failure the runner exits 2, writes the reason to stderr, and emits ok: false with results: null, never an empty list, so a missing credential or an HTTP 402 cannot be read as "there are no such people".
04 同仓库更多 Skills
zero-downtime-cloud-move
Runbook for moving a live app from one cloud to another (Google Cloud to AWS, AWS to Google Cloud, a PaaS to your own account, or one storage provider to another) without downtime. The agent inventories what is really running, builds the new stack next to the old one, keeps a one-line rollback for every phase, proves the new stack with real workloads before any traffic moves, cuts over with a checklist, watches a fixed window after the switch, copies data with an idempotent job whose second pass must report 0 left, and deletes the old cloud last. Use when someone asks to migrate, move or replatform a production app, switch cloud providers, move a backend, database files or object storage, plan a DNS cutover, or decommission an old cloud account.
zayats-ceo-branding
Build an evidence-based CEO positioning paper, Personal Branding Roadmap, and editorial workflow using Marina Zayats's public method. Use for executive branding, LinkedIn ghostwriting briefs, profile strategy, and recurring CEO-approved content planning.
yc-app-roast
YC App Roast: roast a YC or a16z Speedrun application. Use when a founder wants their YC or Speedrun answers scored, asks "what would a YC partner think", wants the weak answers found or rewritten, or wants their draft checked before they submit. Scores every answer 0-10 against public advice from YC and Speedrun partners, names the top issue with the exact words that cost points, and writes a rewrite from the founder's own facts. Works for YC and a16z Speedrun. Not affiliated with Y Combinator or a16z.
yc-app-review
YC App Review: get a YC or a16z Speedrun application reviewed by people you trust. Use when a founder wants friends, mentors or alumni to review their YC or Speedrun application, or asks to share their draft for feedback. Builds a shareable review page (one HTML file, application-portal look) where reviewers leave notes per question and press "Send my notes" to send them back as text, then parses the notes into proposed rewrites. Optional tracked private links per reviewer. Works for YC and a16z Speedrun. Not affiliated with Y Combinator or a16z.
x-growth-playbook
X Growth Playbook. Analyse a person's own X (Twitter) posts and tell them what to post next, with the numbers behind every claim. Compares quote posts vs originals, links vs no links, short vs long, time of day in their timezone, media vs text and the topics they name, reports medians with n, then writes a 10-step plan and the next 7 posts in their voice. Use when the user says "analyse my X posts", "what works on my Twitter", "how do I grow on X", "what should I post next on X", "audit my X account", or shares an X data archive, X API export or a list of their posts. Runs on the user's computer.
workplan
Create, update, or close work plans for multi-step tasks. Use when starting refactors, bug lists, feature work, migrations, or any task with 2+ steps. Also use after auto-compaction to re-orient. Triggers: "workplan", "work plan", "create a plan", "what's the plan", "where was I", or when Claude detects a multi-step task that needs tracking.