job-recruiter-info

Finds recruiter, hiring-manager and careers contacts for one job or company. Use when the user wants to know who to apply to ("find a recruiter", "who recruits for X", "careers email for X"). Not for

SKILL.md

GitHub

Job Recruiter Info

Overview

You are an orchestrator that runs a one-shot recruiter-contact lookup for a single job target. Given a job listing URL or a company name plus an optional role hint, you dispatch the job-contact-finder agent against that target and produce a compact markdown report containing the named recruiter (if any), careers-page email, LinkedIn company page, generic-pattern guesses, apply URL, and a confidence assessment.

This is the standalone counterpart to Phase 6 of /job-search. Use it when the user already has a target job and only needs the contacts — no CV, no scoring, no cover letter, no pipeline state.

When to Use

Looks up apply-target contacts for one job listing or one company, without running the full job-search pipeline.

When NOT to use

  • The user wants to scan many jobs against a CV — use /job-search.
  • The user wants company contacts for B2B outreach (sales, partnerships) — use /outreach-research (full pipeline) or /outreach-enrich-company (deepen an existing card) in mz-biz-outreach.
  • The user wants to draft a cover letter for that role — use the copywrite skill from mz-creative.
  • The user wants to investigate the company itself (reputation, growth, tech) — use /outreach-research for a single-company scan.

Input

  • $ARGUMENTS — Free-text identifying the target. Either:
    • A canonical job listing URL (https://jobs.lever.co/..., https://boards.greenhouse.io/..., LinkedIn job link, etc.), OR
    • A company name plus an optional role hint (e.g. "Acme Corp — Senior Backend Engineer" or just "Acme Corp").
  • Optional modifiers (case-insensitive):
    • region:<value> — geographic hint for disambiguation (e.g. region:DACH, region:EU).
    • language:<value> — language hint when the user expects non-English recruiter contacts (e.g. language:de, language:fr).

If $ARGUMENTS is empty, ask the user for a job URL or a company name before invoking.

Argument Parsing

Extract from $ARGUMENTS:

  • target — the listing URL or company name (everything not matching a parameter pattern).
  • role_hint — optional, the role description if combined with a company name (split on —, -, :, or ,).
  • region — optional, from region:<value> (default: derived from the listing URL host or the company domain TLD).
  • language — optional, from language:<value> (default: en).

Directory Structure

  • State — <PROJECT_ROOT>/.mz/task/<task_name>/state.md. Source of truth across phases.
  • Output — <PROJECT_ROOT>/.mz/outreach/<run_name>/ (called <RUN_DIR> in phase files) holds target.json (parsed target), contacts.json (raw agent output), and the final <YYYY_MM_DD>_recruiter_info_<slug>.md report.

task_name follows <YYYY_MM_DD>_job_recruiter_info_<slug>; run_name is <YYYY_MM_DD>_job_recruiter_info_<slug> (same leading-date convention; max 40 chars). Same-day collisions append _v2, _v3.

Constants

  • PROJECT_ROOT: resolved once in setup with echo "${MZ_MEMORY_ROOT:-$(git rev-parse --show-toplevel 2>/dev/null || pwd)}"; expand <PROJECT_ROOT> to it in every command and agent prompt.

Core Process

Phase Overview

# Phase Agent(s) Details
0 Setup + target parsing — (orchestrator) Inline below
1 Contact lookup job-contact-finder phases/lookup_and_report.md
2 Compact markdown report — (orchestrator) phases/lookup_and_report.md

Read phases/lookup_and_report.md when you reach Phase 1.

Phase 0: Setup + target parsing

Resolve PROJECT_ROOT with the command in Constants, before any cd or worktree switch, and use it for every .mz/ path below.

Parse arguments. Derive:

  • task_name (state dir) = <YYYY_MM_DD>_job_recruiter_info_<slug> where <slug> is a snake_case summary of the target (max 20 chars, derived from company name or listing host).
  • run_name (outreach output dir) = <YYYY_MM_DD>_job_recruiter_info_<slug> (max 40 chars).
mkdir -p "<PROJECT_ROOT>/.mz/task/<task_name>"
mkdir -p "<PROJECT_ROOT>/.mz/outreach/<run_name>"

Detect input shape:

  • URL — target matches ^https?://. WebFetch the page once to extract: company, role_title, location_string, summary_snippet. If WebFetch fails, fall back to the URL host as the company hint and surface the failure as a DONE_WITH_CONCERNS note in the final report.
  • Company name (with optional role hint) — Use target as company; use role_hint as role_title if provided. Derive the company domain via WebSearch ("<company>" official site); first authoritative result is the domain.

Write <PROJECT_ROOT>/.mz/outreach/<run_name>/target.json:

{
  "input_type": "url" or "company",
  "input_raw": "<original argument>",
  "company": "<resolved company name>",
  "role_title": "<resolved role title or null>",
  "listing_url": "<resolved URL or null>",
  "company_domain": "<resolved domain or null>",
  "location_string": "<from listing or null>",
  "summary_snippet": "<from listing or null>",
  "region": "<from arg or inferred or null>",
  "language": "<from arg or 'en'>"
}

Write <PROJECT_ROOT>/.mz/task/<task_name>/state.md:

schema_version: 2
Status: running
Phase: 0
Started: <ISO timestamp>
phase_complete: false
what_remains: []
Target: <input_raw>
InputType: <url|company>
Company: <resolved company>
RoleTitle: <resolved role or null>
RunName: <run_name>

After setup completes, read phases/lookup_and_report.md and proceed to Phase 1.

Red Flags

  • You fabricated a recruiter name or email that did not appear in any verified source.
  • You included a noreply@, no-reply@, donotreply@, mailer-daemon@, or postmaster@ address in the report.
  • Marked confidence: high without a named recruiter AND a verified email (verified means seen on the listing, careers page, or a named LinkedIn handle).
  • Followed a LinkedIn profile URL deep into the platform (only the search-results page is allowed per agent rules).
  • Report lives in chat instead of .mz/outreach/<run_name>/<YYYY_MM_DD>_recruiter_info_<slug>.md.

Verification

Before completing, output a visible block showing: target (company + role if known), confidence level (high/medium/low), whether a named recruiter was found, careers-page email status (verified/guessed/none), and the absolute path of the report. Confirm both contacts.json and the report file exist on disk.

State Management

State persists to .mz/task/<task_name>/state.md under the project root, ${MZ_MEMORY_ROOT:-$(git rev-parse --show-toplevel 2>/dev/null || pwd)}, resolved once at the start of the run. Schema is v2: line 1 is schema_version: 2; line 2 is Status: with one of pending, running, blocked, complete, aborted_by_user or failed, lowercase and unquoted; then Phase, Started, phase_complete (boolean) and what_remains (YAML list of strings), with skill-specific keys below them. Milestones go in Phase; the reason for blocked or failed goes in what_remains. Set phase_complete: false on phase entry and true once the phase's artifacts are written and its gates pass. Refresh what_remains on every phase transition; it is [] when Status is complete. Write the file after every phase and before every gate, and re-read it after compaction. Record decisions worth recalling under a heading that reads exactly ## Decisions, placed before ## Notes; mz-memory harvests them only once Status is complete. Under a heading that reads exactly ## Lessons, add one line whenever a review finding is accepted and fixed, a step fails, gate feedback corrects a mistake, or an approved approach is reversed, in the form - (review|failure|correction|change) WHEN <situation> → <what to do> — WHY <cause>, concrete and under about 40 words; mz-memory harvests lessons whatever the Status. Treat a legacy completed as complete. On reading a schema_version: 1 or unversioned file, put schema_version: 2 on line 1 and Status on line 2, add the missing keys, and log the upgrade. Leave a file with a higher schema_version untouched: ask whether to Restart (archive it), start New (_v2) or Abort, and stop if nobody can answer.

Resume Support

Before creating anything in Phase 0, check if <PROJECT_ROOT>/.mz/task/<task_name>/state.md exists for the resolved task_name. If it does, read the Phase field and resume from the next incomplete phase. All phases are idempotent — re-running overwrites output files.

Error Handling

  • Agent returns empty: append an Errors: bullet in state.md, write a report that explicitly says "no contacts surfaced" with the ZERO RESULTS VERIFIED or ZERO RESULTS UNVERIFIED disclosure tag from the agent.
  • Agent returns STATUS: NEEDS_CONTEXT: invoke AskUserQuestion to fill the missing field (typically the company domain), then re-dispatch.
  • Agent returns STATUS: BLOCKED: stop the pipeline, set Status: blocked with the cause in what_remains, surface the cause to the user.
  • Never fabricate data — incomplete results are better than false results.
  • Listing URL was provided but unreachable: degrade to input_type: company, surface the WebFetch failure in the final report.