Exec Recruiter
bgolubovski/exec-recruiter
Recrutador executivo pessoal para lideranças em produto, sem retainer.
Ver no GitHubPlugin de recrutamento pessoal para buscas de liderança em produto (CPO, VP Product, Head of Product). Automatiza o pipeline completo: sourcing diário em LinkedIn e ATSs, triagem com score de fit, geração de CV tailorizado, rascunho de outreach e prep para entrevistas com pesquisa de entrevistador.
Instalação
npx convez-ai bgolubovski/exec-recruiter1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.
Métricas
Casos de uso
- Sourcing automatizado de oportunidades em LinkedIn e plataformas ATS
- Triagem e scoring de vagas contra critérios de fit customizáveis
- Geração de CV tailorizado e rascunho de outreach por canal
- Prep estruturado para entrevistas com pesquisa de entrevistador e company intel
- Debrief pós-entrevista com estrutura de sinais e próximos passos
Veredito

Solução sofisticada e bem pensada para executivos em transição de carreira, com foco em approval-gating e controle manual em etapas críticas. Requer config inicial robusta e disciplina no uso, mas entrega workflows de nível recruiter sem custos de retainer.
README
A tradução para o português deste README ainda está na fila.
exec-recruiter
A personal executive recruiter for Product Leadership job hunts (CPO, VP/SVP Product, Head of Product at scale). The plugin runs the full pipeline: daily sourcing, fit-scored triage, JD-tailored CV generation, contact-aware outreach drafts, approval-gated sends, meeting / interview prep with interviewer research, and post-call debrief. Built for senior product operators who want a recruiter-grade workflow without burning a recruiter retainer.
What it does
Five phases, all approval-gated.
Phase 1: Source and rank. Sweeps LinkedIn AND major ATS domains (Greenhouse, Lever, Workday, Ashby, iCIMS, Jobvite, BambooHR, SmartRecruiters, Workable, Teamtailor) via Google site: operator. Pulls full JDs via the apply-link redirect to the underlying ATS or directly from the ATS, scores each role 1-10 against your fit criteria, dedupes against your existing dashboard, and produces a Tier 1 / Tier 2 / Skip greenlight summary.
Phase 2: Packet build. For each role you greenlight, identifies the right contact (job poster, hiring manager, 1st-degree, 2nd-degree mutuals), tailors the CV to the JD using a discipline that keeps identity stable (only the adjacency sentence, one or two highlight cells, and per-role italic context lines change), and drafts channel-appropriate outreach (Connect+Note, longer DM if Open Profile available, InMail as last resort).
Phase 3: Send. Walks you through send one role at a time. The plugin opens the right LinkedIn surface, types the message into the field, and stops. You click Send. No autonomous sending.
Phase 4: Meeting / interview prep. When a contact accepts, replies, or books a meeting, the plugin builds a prep packet for the call. Researches the named interviewer (LinkedIn, TheOrg, Crunchbase, web), refreshes company intel, generates the scope-disambiguation question, first-90-days narrative (with branches if scope is ambiguous), questions to ask, one-line answers to likely questions back, comp posture, friction flags, and a wrap-up signal checklist. Saves three artifacts per call: full markdown, interactive HTML with TOC and Call-mode toggle, and a scannable one-page card. Supports rounds: screen, working-session, panel, final, exec, reference-check-prep. --deep flag triggers maximum-depth company intel.
Phase 5: Post-call debrief. After a Phase-4-prepped call, the user pastes their notes. The plugin parses what was said, captures scope / comp / process / friction signals, structures a debrief artifact, updates the dashboard, and identifies the next action (prep next round, send thank-you and wait, send requested materials, passive watch, mark rejected, or escalate to a different contact). Prompts for memory writes when patterns emerge that contradict prior assumptions.
Configurable surface
Seven config files in config/ define the user-tunable parts. Edit them at install time, then re-run.
icp.yaml- target role titles, search keywords, seniority, sector preferencesgeo-rules.yaml- default location, opt-in regions, geo-conditional CV header stringsscoring.yaml- dimension weights for the 1-10 fit score (stage, domain, scope, comp, geo, founder, etc.)channels.yaml- DM / Connect+Note / InMail order, credit-preservation rules, character limitsfollow-up.yaml- Day-5 / Day-7 / Day-10 thresholds for bumps and final-actionoutreach-style.yaml- tone, banned punctuation (em/en dashes), sign-off, languageats-sources.yaml- the 10 ATS domains scanned in Phase 1B, with per-ATS liveness checks and SPA-render routing
CV baseline
The plugin does NOT ship with a CV baseline. Place your own master CV at templates/cv-baseline.docx (any executive product CV will do). The tailoring discipline preserves your identity blocks (name, headline scaffold, work history dates, education) and only swaps:
- The headline-positioning fragment
- The adjacency sentence at the end of the executive summary
- One or two career-highlight cells (out of six)
- The italic context line per past role
- The geo-conditional header
This lets you ship a JD-tailored CV without the false-positive rebuild risk. The structural expectations are documented in scripts/tailor_cv.py and agents/cv-tailor.md.
Note: the CV filename convention currently includes the operator's name (e.g. Operator-FirstName-CV-<slug>.docx). To use this plugin as-is, customize the prefix in scripts/tailor_cv.py and agents/cv-tailor.md to match your own name, OR wait for the planned config/operator.yaml parameterization in a future version.
Skills
| Skill | Triggers on |
|---|---|
phase-1-sweep | "run today's sweep", "Phase 1", "find new roles" |
phase-2-packet | "build packet for X", "tailor CV for X", "Phase 2" |
phase-3-send | "send X", "Phase 3", "ready to send" |
phase-4-prep | "prep for X", "build prep packet for X", "interview prep X", "/prep " |
phase-5-debrief | "debrief X", "log my call with X", "/debrief " |
score-role | "score this JD", "is this role a fit" |
daily-bump | "what bumps today", "Day-5 follow-ups" |
dashboard-status | "where do I stand", "today's actions" |
Agents
jd-extractor- pulls full JDs via the LinkedIn-to-ATS redirect URL extraction trick (Teamtailor, Workable, Greenhouse, Lever, Ashby, company-careers pages). Solves the LinkedIn-skeleton-view problem.ats-google-sweep- searches the 10 major ATS domains via Googlesite:operator. Liveness-checks every URL before fetching, routes SPA-based ATSs (Ashby, Workday) through Chrome render.cv-tailor- applies the tailoring discipline rules to the CV docx, regenerates PDF via libreoffice.outreach-drafter- generates Connect+Notes (300 char Premium / 280 standard), longer DMs, and InMail variants per channel preferences.meeting-prep-researcher- researches a named interviewer (LinkedIn, TheOrg, Crunchbase, web) and refreshes company intel for Phase 4 prep packets.
Templates
templates/dashboard.html- rich live dashboard with cohort views, score histogram, domain breakdown chart, follow-up tracker, inbound-signal detection. Ships with emptyROLES = []; phase-1/2/3/4/5 workflows populate it over time.templates/jd.md.template- JD-capture template used by Phase 2templates/outreach.md.template- outreach drafts template used by Phase 2templates/prep-meeting.md.template- full markdown prep packet template used by Phase 4templates/prep-meeting.html.template- interactive HTML prep packet (light-mode UI, sticky TOC, sticky sidebar, view toggle Full / Call mode, persistent checkboxes, print-friendly CSS) used by Phase 4templates/prep-onepager.md.template- scannable one-page prep card used by Phase 4
Required connectors
| Category | Required for | Options |
|---|---|---|
| Browser | LinkedIn navigation, JD extraction, send-flow positioning | Claude in Chrome MCP |
| Shell | File ops, libreoffice for PDF generation | Workspace bash (built-in) |
| Web search | ATS-direct sourcing, interviewer research, company intel refresh | WebSearch tool (built-in) |
| Web fetch | LinkedIn / ATS / TheOrg / Crunchbase fetches | mcp__workspace__web_fetch (built-in) |
| Contact enrichment (optional) | Better contact identification when 1st/2nd-degree path is unclear | Apollo MCP |
| Calendar (optional) | Scheduled task setup for daily 9am sweep | Cowork scheduled tasks (built-in) |
Setup
- Install the plugin
- Edit the seven YAML files in
config/to match your ICP, geo rules, scoring weights, channel preferences, follow-up windows, outreach style, and ATS sources - Drop your own master CV in
templates/cv-baseline.docx(keep the section structure: name + headline + tailored fragment, executive summary with adjacency sentence at end, six-cell career-highlight grid, experience with italic context line per role, core competencies, thought leadership, education) - Customize the CV filename prefix in
scripts/tailor_cv.py+agents/cv-tailor.mdfromOperator-FirstName-CV-to match your own name, OR rename consistently - Initialize a working dashboard: copy
templates/dashboard.htmlto your project folder - Customize the operator-personal placeholders in
templates/prep-meeting.html.template(anchor names, comp range, location, etc.) OR let Phase 4 populate them at runtime from your actual project context - Trigger your first sweep with
Phase 1or "run today's sweep" - Optional: set up the daily 9am scheduled sweep via
mcp__scheduled-tasks__create_scheduled_task(thephase-1-sweepskill includes the command)
Memory and state
The plugin keeps no persistent state of its own. Per-role packets live under outreach/{slug}/ in your project folder. Status, follow-up windows, and inbound signals are tracked in dashboard.html. Loss of plugin state never loses your sourcing history.
Send safety
No skill in this plugin sends a LinkedIn message, an email, or an application autonomously. Every send is a two-step: the plugin types into the right field and stops; you click Send. The Phase 3 skill explicitly enumerates the approval gate before each action. Phase 4 generates prep material only; it never sends outreach. Phase 5 captures debrief notes only; it updates the dashboard but never sends follow-up messages.
Versioning
See CHANGELOG.md for the full version history. Current version: 0.5.1.