---
name: recruiter-candidate-pipeline
description: Daily inbox-to-pipeline update for a recruiter, a waiting-on-you list, and scheduling and status drafts. Never sends, books, advances, or rejects a candidate without the owner's Allow.
---

# Recruiter Candidate Pipeline: keep the table current, draft the replies

Use this as a weekday morning routine (for example, 8:00). Work only with candidate data the owner is allowed to process. Do not mass-message, scrape contact details, or make hiring decisions.

## Inputs
- Inbox (connector), ideally a label such as Candidates
- Pipeline table: Notion database or Google Sheet with columns for name, role, stage, last contact, next step
- Calendar (connector) for interviews
- Waiting rule: default 3 days since a candidate's last message with no reply
- Open roles and their hiring managers

## Steps
1. Read new candidate emails since the last run. Add or update one row per candidate. Write "missing" for unknown fields. Do not delete rows.
2. List candidates waiting on the owner past the waiting rule, oldest first.
3. For each waiting candidate, draft a short, honest reply or status update. Save as drafts.
4. For candidates who need an interview, find three slots when the owner and the hiring manager are both free and draft an email offering them. Do not create the event until the candidate picks and the owner allows it.
5. On Fridays, count candidates by stage per role. Write "missing" for anything you can't count.

## Privacy rules
- Don't copy CVs or personal data into places the owner didn't name.
- Don't add protected characteristics to the table or summaries.
- Summaries of a CV must quote or point to the CV for each claim.

## Hard stops
- Never send, book, advance, or reject without the owner's Allow.
- If a site asks for a login, CAPTCHA, or code, hand it back to the owner.

## Output
1. Pipeline changes made
2. Waiting-on-you list with drafts
3. Scheduling drafts
4. (Fridays) stage counts per role