galiprandi/job-seeker / .agents/skills/referrals
referrals
VerifiedDiscovers internal contacts (1st/2nd degree connections, university alumni, ex-colleagues) and recruiters at target companies, stages personalized referral requests or outreach DMs in DB, and applies dynamic ATS micro-alignment (JD-to-CV tailoring).
Warm Sourcing & Referrals
Trigger
Keyword: referrals (or variants: "warm sourcing", "buscar contactos", "solicitar referido")
The user says referrals or launches warm sourcing for a target company/role. Also executed as step 0 of the apply and targets flows to maximize conversion.
Flow
0. Pre-flight
- Verify active browser session (see AGENTS.md "Browser session"):
node .agents/skills/browser-automation/scripts/browser.js open <url> --headed(Gold Rule 5) if session closed - Load profile, university background, past companies, and job preferences from Postgres DB:
node scripts/db.js "SELECT data->'profile' AS profile, data->'job_preferences' AS prefs, data->'style_profile' AS style FROM users WHERE id = <user_id>" - Load strategy (see AGENTS.md "Strategy levels"):
Respect:node scripts/db.js "SELECT data->'strategy' AS strategy FROM users WHERE id = <user_id>"cold_outreach(gates the recruiter-outreach branch in step 3). Ifreferralsis not insources_active, the flow should not run standalone — when invoked as step 0 ofapply/targets, those flows handle the gate. - Load active preferences (see
memoryskill):node scripts/db.js "SELECT category, key, value, confidence, source FROM preferences WHERE user_id = <user_id> AND status = 'active' ORDER BY category, key"
1. Warm Contact & Recruiter Discovery
For a target company and role:
# Automated discovery script
node scripts/linkedin-warm-sourcing.js --company "<Company>" --role "<Role>" --json
The script searches for:
- 1st & 2nd degree connections currently working at
<Company> - University alumni (matching institutions from
users.data.profile.education) - Ex-colleagues (matching past employers from
users.data.profile.experience) - Recruiters & Hiring Managers assigned to the role/company
2. Referral Request Staging (Highest Conversion — Strategy #1)
If an internal contact, alumni, or ex-colleague is found:
- Do NOT submit a cold application immediately. A referral yields a 40% hire rate vs 2-3% for cold Easy Apply.
- Draft a personalized referral request message:
- Must pass Gold Rule 7 (Anti-LLM Checklist): no em-dashes, no bullet points, conversational tone, max 2 short paragraphs, natural mention of shared background (alumni/ex-colleague/interest).
- Tone: polite, non-demanding, asking for team insights or guidance on applying.
- Stage the draft in DB:
node scripts/db.js "INSERT INTO messages (user_id, channel, direction, sender, subject, body, draft, status, received_at, data) VALUES (<user_id>, 'linkedin', 'outbound', '<contact_name>', 'Solicitud de referido / consulta sobre equipo', '', '<draft_text>', 'draft', NOW(), '{\"category\": \"referral_request\", \"company\": \"<Company>\", \"vanity\": \"<vanity>\"}'::jsonb)" --write - Register or update pipeline card in stage
discovered:node scripts/pipeline.js --move <id> discovered
3. Recruiter Outreach Staging (Multi-channel Combo — Strategy #4)
If NO internal referral path exists:
- Gate: if
strategy.cold_outreach = false→ skip this step. Proceed to step 4 (ATS micro-alignment) and cold apply only. - Extract the Recruiter / Hiring Manager profile vanity or email.
- Prepare a personalized recruiter DM outreach draft (3-4 lines: trigger + credibility anchor + clear ask).
- Stage the draft in DB (
messagestable withcategory: recruiter_outreach). - Proceed to cold postulation via
applyortargetswhile keeping the recruiter outreach staged for user approval (surfaced bynewsflow).
4. Dynamic ATS Micro-Alignment (JD-to-CV Tailoring)
Before submitting an application via ATS or email:
- Extract top 5 technical & domain keywords from the target Job Description (e.g.,
LangChain,System Architecture,PyTorch,Technical Leadership). - Compare against
users.data.profile.skillsandusers.data.cv_markdown. - Highlight matching achievements in the top summary/highlights of the CV markdown.
- Generate the micro-aligned PDF CV using
scripts/generate-cv.jsbefore submitting:node scripts/generate-cv.js --output assets/cv_tailored_<company>.pdf
5. Presentation & Summary
Present the warm sourcing results to the user:
- Internal contacts / Alumni found: list with profile URLs and proposed referral draft.
- Recruiters found: list with proposed DM outreach draft.
- Tailored CV generated: link to tailored PDF.
Dependencies
- Depends on
onboarding(DB to register) - Depends on
profile(education & past experience data for alumni/ex-colleague matching) - Integrated into
applyandtargetsflows
Script reference
scripts/linkedin-invite.js -- Send connection requests
Navigates to /preload/custom-invite/?vanityName=<vanity>, clicks "Send without a note". Anti-ban delay of 3s between invites.
# Invite one or more vanities
node scripts/linkedin-invite.js <vanity-name>
# Invite multiple
node scripts/linkedin-invite.js vanity1 vanity2 vanity3
# Search + invite in one command (pipe search -> invite)
node scripts/linkedin-invite.js --from-search '"<Role>" "hiring" <Region>'
Flags: --from-search "<keywords>" (searches and invites all found)
Exit codes: 0 = at least one sent, 1 = all failed, 2 = error
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack." | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |