agentsclimarketplace

Vertical hr recruiting

Skill avelikiy/great_cto/skills/vertical-hr-recruiting

Domain-knowledge primer for the HR & recruiting vertical (ATS, onboarding, workforce scheduling, engagement). Applied by architect/pm during spec authoring so they aren't naive about hiring pipelines, the admitted offer→onboard data-carry gap, EEO/I-9 compliance, and shift-coverage rules. Stops the four products from being specced as generic CRUD when the domain has hard legal and workflow constraints.From its SKILL.md

Install
npx -y skills add avelikiy/great_cto --skill vertical-hr-recruiting

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

SKILL.md

8.7 KB, ~1.9k tokens by cl100k_base, as published. Nobody here has run it

Vertical: HR & recruiting — don't spec it naive

The incumbents (Workable, BambooHR, Greenhouse, Lever, Zoho Recruit, Manatal) have trained buyers to expect a hiring pipeline that works. A generic CRUD app fails the moment it meets EEO law, I-9 timing, or the offer→onboard handoff. Read this before speccing any of the four products — the domain has hard constraints, not just forms.

1. Domain vocabulary

  • ATS — applicant tracking system; the system of record for hiring.
  • Requisition (req) — an approved open role. Hiring happens against a req, not in a vacuum. Reqs have an approval workflow (hiring manager → finance/HR).
  • Pipeline stagesapplied → screen → interview → offer → hired (plus rejected / withdrawn). Stages are configurable per req — engineering and sales hire differently.
  • Candidate vs applicant — an applicant applied to a specific req; a candidate is a person in your talent pool who may map to many applications over time. Don't conflate them.
  • Sourcing — proactively finding candidates (vs inbound applications).
  • Structured interview + scorecard — pre-defined questions + a rubric each interviewer scores. Reduces bias and legal exposure vs freeform notes.
  • Offer letter — formal terms; triggers the offer→onboard transition on acceptance.
  • EEO data — voluntary race/gender/veteran/disability self-ID, collected for reporting, segregated from hiring decisions (see §2).
  • I-9 + E-Verify — employment eligibility verification; strict 3-day timing (§2).
  • Onboarding checklist — tasks a new hire/employer must complete before/at start.
  • time-to-hire / time-to-fill — core recruiting metrics (hire = offer accepted by a candidate; fill = req closed). Different denominators; report both correctly.
  • Hourly vs salaried — drives scheduling, overtime (FLSA), and pay rules.
  • Shift swap / coverage — hourly workers trade shifts; coverage rules say a slot can't go unstaffed below a threshold.
  • eNPS — employee Net Promoter Score; the headline engagement-survey metric.

2. Non-obvious domain rules

  • The offer→onboard handoff is the admitted gap. Incumbents openly concede onboarding is unsolved: data gets re-entered between the ATS and the HR/onboarding system. The whole onboarding product wedge is carry the candidate's data forward — zero re-entry.
  • EEO/OFCCP data must be collected but kept OUT of the hiring-decision view. Mixing self-identified race/gender into the screen/interview UI is an anti-discrimination liability. Store it segregated; surface it only in aggregate compliance reports.
  • Structured scorecards reduce bias and legal risk. They create a defensible, consistent record. Freeform-only interview notes are a disparate-impact landmine.
  • I-9 has strict timing: Section 1 by the employee's first day, Section 2 (employer review of documents) within 3 business days of start. Onboarding tasks tied to I-9 carry a hard deadline, not a soft reminder.
  • Workforce scheduling needs coverage rules + labor compliance. A schedule isn't valid just because slots are filled — it must respect minimum coverage, overtime (FLSA), predictive-scheduling / fair-workweek laws (advance-notice in some jurisdictions), and break rules.
  • Reqs gate hiring. No offers without an approved req; req approval is a real workflow with a budget/headcount check, not a checkbox.

3. What a naive build gets wrong

  • Hardcoded pipeline stages. A fixed applied→hired enum breaks the first time a customer wants a take-home or panel stage. Stages must be configurable per req.
  • EEO data in the candidate decision view. Putting self-ID fields on the candidate card the hiring manager sees is a legal-risk bug, not a UX choice. Segregate it.
  • Onboarding that re-enters candidate data. Rebuilding name/email/role/comp from scratch is the gap incumbents have. If onboarding starts from a blank form, you built the incumbent's weakness, not our wedge.
  • I-9 modeled as a generic checklist item. Without the 3-business-day deadline and Section 1 / Section 2 split, it's non-compliant.
  • Scheduling as a calendar. Drag-and-drop shifts with no coverage minimum, no overtime flag, and no swap-approval flow is a toy, not a workforce tool.

4. Must-model entities

EntityMust includeWhy
Requisitionapproval state, headcount, stage confighiring is per-req; stages vary
Candidatestage history, scorecards, EEO (segregated store)audit trail + bias defense; EEO must not leak into decision view
Offer → Onboardingcarries candidate data forward (no re-entry)this IS the wedge; the handoff is the gap
OnboardingTaskdeadline field (I-9 3-day timing)compliance is time-bound, not soft
Shiftcoverage rules, overtime flag, swap/approval statea schedule must be valid, not just full

5. Per-product notes

productarchetypewedgethe one domain thing
atscrudcrowded market — needs a sharp angle, not "another tracker"configurable-per-req pipeline stages + structured scorecards; don't ship a generic Kanban
onboardingcrudthe offer→onboard data-carry gap incumbents admit is unsolvedcarry candidate data forward with zero re-entry; I-9 tasks with 3-day deadlines
workforce-schedulingbookingshift scheduling for hourly teamscoverage rules + overtime + swap/coverage approval, not a bare calendar
engagementcrmengagement surveys for SMBeNPS as the headline metric; anonymity threshold so small teams can't deanonymize responses

onboarding is the differentiated bet (solve the admitted gap). ats is the crowded one — architect must name the sharp angle in the ARCH doc or it's dead on arrival.

6. Compliance (light)

Flag these in the ARCH doc; route AI-screening to a reviewer.

  • EEO / OFCCP — collect voluntary self-ID; aggregate reporting; never in the decision path.
  • I-9 / E-Verify — Section 1 by day one, Section 2 within 3 business days; model the deadline.
  • Anti-discrimination — disparate impact in screening. If any product uses AI to screen/rank candidates, flag it to hr-ai-reviewer (or the AI-security reviewer) before senior-dev starts — automated screening is a high-risk surface.
  • Ban-the-box — many jurisdictions forbid asking criminal history before an offer. Don't put it on the application form by default.
  • FLSA / overtime — scheduling must compute overtime for hourly workers; respect fair-workweek / predictive-scheduling laws where they apply.
  • Data retention — candidate/applicant records have minimum retention (e.g. EEOC ~1 year) and deletion obligations (GDPR/CCPA right-to-erasure). State a retention policy.

Output

When applied, contribute a Domain constraints block to the architecture doc:

## Domain constraints (HR/recruiting)
- pipeline: stages configurable per req (not a fixed enum)
- EEO data: segregated store, excluded from decision view
- offer→onboard: candidate data carried forward, zero re-entry
- I-9: Section 2 deadline = start + 3 business days (modeled)
- scheduling: coverage rule = <min staffed>, overtime flag, swap-approval
- AI screening: <none | flagged to hr-ai-reviewer>
- retention: <policy + jurisdiction>

Cross-references

  • [[vertical-onboarding]] — the import-first onboarding funnel; the offer→onboard data-carry here is the import that feeds first-run for the onboarding product.
  • [[migration-ready-schema]] — model Candidate/Offer/Onboarding so data carries forward cleanly (the no-re-entry wedge depends on it).
  • [[lifecycle-messaging]] — candidate-stage and onboarding-task notifications (offer sent, I-9 due, shift posted, survey open).

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Gives 0 of the 12 instructions most hr recruiting skills give in ~1.9k tokens

Counted across 356 of the 357 authors here whose files we hold, read 2026-08-07

  • Quantify achievements with specific metricsin 14 of 356, across 6 files
  • Keep the resume under two pagesin 14 of 356, across 6 files
  • Request the full job description if not providedin 12 of 356, across 4 files
  • Extract keywords and prioritize job requirementsin 12 of 356, across 4 files
  • Stop and ask for clarification if required inputs are missingin 12 of 356, across 5 files
  • Map candidate experience to job requirementsin 11 of 356, across 3 files
  • Ask if the user wants adjustmentsin 11 of 356, across 3 files
  • Provide strengths and gap analysis after the resumein 10 of 356, across 2 files
  • Request candidate background details if not providedin 10 of 356, across 2 files
  • Format experience bullets as action verb plus resultin 10 of 356, across 2 files
  • Ask for missing inputs before startingin 10 of 356, across 9 files
  • Use exact job description terminologyin 9 of 356, across 1 file

Said here and by no other author read

  • make pipeline stages configurable per req
  • segregate EEO data from the hiring decision view
  • carry candidate data forward during onboarding
  • model I-9 tasks with a three business day deadline
  • enforce scheduling coverage rules and overtime flags
  • require shift swap approvals

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 325,949. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.