skaills
Lever integration

Add assessment depth to your Lever

skaills plugs into Lever via the Data API and Postings Webhooks. No changes on your side. No disruption to your sourcers' flow. Every candidate entering the pipeline advances with a deep competency profile and a cited AI score.

Three ways to integrate

Pick the depth that fits your needs — start light, grow when ready.

Outbound — skaills enriches the candidate in Lever

Planned
skaills pushes results into Lever via the Data API as opportunity profile fields and structured notes. Examples: candidate completes a skaills evaluation → multi-dimensional score appears as a field on their Lever profile, PDF as file attachment. Psychometric score posted as a structured note on the candidate feed. Advance recommendation visible on the Lever pipeline card.
~2 weeks for the first connector. No Lever-side custom work if the Data API key is enabled.

Bidirectional — Lever drives, skaills assesses

Planned
Lever is the talent system of record; skaills plugs in as a specialty assessment provider. Lever Postings Webhooks notify skaills when a candidate advances to a configured stage (e.g. "Evaluation"). skaills generates an evaluation link, sends it to the candidate, computes the multi-dimensional score, and pushes the result back to Lever via the Data API. The sourcer/recruiter sees everything in their usual Lever console.
~4-6 weeks. Requires Lever-side configuration (Webhooks, custom fields for our scores, "Evaluation" stage).

Lever Partner Ecosystem — skaills as certified integration

Roadmap
Published in the Lever Partner Ecosystem as a certified assessment integration. Category: Assessments + Talent Intelligence. Analogues today: Codility, HackerRank, Plum, Criteria Corp.
4-6 months including Lever Partner certification, security review, listing.

What this looks like in practice

Concrete flows triggered by events in Lever.

Trigger in LeverWhat skaills doesWhat flows back to Lever
Sourcer adds a passive candidate to Lever via Chrome extensionskaills runs 12-dimension AI scoring of the CV/LinkedIn with cited rationaleOverall score + per-dimension breakdown + PDF appear on the candidate profile in Lever before the first nurture touch
Candidate responds positively to a Lever Nurture sequence and advances to "Evaluation"skaills generates an evaluation link tailored to the posting, sends to candidate, computes psychometric + competency scoreScore, ranking vs. other candidates for the posting, and structured report come back to Lever as a completed feedback form
Recruiter identifies a re-engagement talent pool in Leverskaills re-evaluates the pool against a new posting, identifies top fits, suggests personalized re-engagement messagePrioritized list with scores + suggested messages appear as tags and note in Lever; sourcer sends nurture sequence
Candidate is hired in Leverskaills auto-creates a 12/24/36-month Personal Growth Plan from the pre-hire evaluationPGP lives in skaills (post-hire Transformative layer); Lever closes the opportunity with PGP link in notes

Technical architecture

The integration uses Lever's public APIs: Data API (read/write — opportunities, postings, stages, feedback forms), Postings Webhooks (stage-change, archived, hired notifications). Authentication via a dedicated Data API key from your Lever account with minimum scopes (opportunities: read/write; postings: read; feedback forms: write). Data via HTTPS (TLS 1.2+). Webhooks land at a dedicated skaills Cloud Function (`leverWebhook`) with signature verification. The API key lives in Secret Manager, never in plain text. All skaills-side logic is serverless.

Implementation timeline

1
Week 1

Kickoff and scope

Define which postings are in scope for the pilot, which pipeline stage triggers the skaills evaluation, which custom fields will be created in Lever to receive scores, and which sourcers/recruiters participate. Identify the project sponsor, the customer-side Lever admin and the skaills team.
Deliverable:Scope document signed + Lever Data API key issued with minimum scopes
2
Week 2

Technical setup

Create custom fields in Lever (overall skaills Score, per-dimension breakdown, report link). Configure the Webhook pointing at our Cloud Function. First end-to-end test with a synthetic opportunity: stage change → webhook → skaills generates link → evaluation completes → score injects into Lever.
Deliverable:E2E test passing with synthetic opportunity; custom fields visible on Lever profiles
3
Week 3

Controlled pilot

Deploy to one real sourcing/recruiting team, with 10-20 real candidates flowing through. We monitor: completion rate, latency, perceived score quality, lift in advance decision, impact on ongoing nurture. 15-minute daily standup.
Deliverable:Pilot report with real metrics and sourcer/recruiter feedback
4
Week 4

Adjustments and rollout

Apply adjustments based on feedback. Internal documentation for your sourcers/recruiters: how to read the skaills report, how to use the score in nurture, where to verify detail. Handoff session with your Lever admin + Recruiting Ops team. Gradual extension to more postings and teams.
Deliverable:Operational runbook + integration running for all in-scope postings

For Lever Partner Ecosystem listing, add another 4-6 months for Lever Partner certification and security review.

Security & compliance

skaills operates under least-privilege principles. The Lever Data API key we use has only the necessary scopes (opportunities: read/write; postings: read; feedback forms: write) — never blanket access to your account. The key is stored in Secret Manager. Signature verification on every inbound Webhook. Sensitive candidate data encrypted in transit (TLS 1.2+) and at rest. skaills doesn't export PII outside the region configured for your tenant. We comply with GDPR requirements — including right-to-erasure and portability. Our SOC 2 posture is inherited from Firebase/GCP; direct SOC 2 Type II certification is on the roadmap.

Ready to add the transformation layer to your Lever?

We'll scope the integration, set up the technical connection, and have you running in weeks — not months.