Jobs Signal

/jobs-signal — Hiring-intent signals from LinkedIn jobs

Companies that are hiring are companies that are spending. A target account hiring 5 RevOps roles signals revenue-tooling intent. A whole ABM list hiring AI engineers signals a market shift. This skill turns LinkedIn jobs data into structured signals for ABM and niche-signal workflows.

Imported via: /steal analysis 2026-05-01 (.claude/discovery/0526-apify-linkedin-actors-steal-analysis.md). Closed the only direct gap on the Apify-actor list — there was no existing capability for hiring-intent signals.

When to use

Run this skill when

• You have an ABM target list and want intent signals layered on
• You're researching a niche and want to know which companies are actively building in it
• A client wants competitive insight into who's hiring for [role] in [vertical]
• Pre-discovery prep wants "is this prospect hiring? for what?" as account context

Do NOT use when

• You need candidate-side data (resumes, profiles for recruiting) — wrong tool
• You need general job-board aggregation (Indeed, Glassdoor) — different surface
• You need company firmographics — use /company-context
• You need contact data at companies — use /clay-search or /deepline-enrich

Two modes — voice-locked

Mode 1 — Bulk scan (default)

Actor: valig/linkedin-jobs-scraper Cost: $0.32/1k jobs (DIAMOND tier) → $0.40/1k (FREE tier) Use when: scanning many companies for hiring activity, or running market-wide role queries

The cheapest reliable jobs scraper at 99.7% success rate, 4.98★. Returns role title, location, posting date, company URL, job description.

Mode 2 — Enriched (recruiter contact + AI enrichments)

Actor: fantastic-jobs/advanced-linkedin-job-search-api Cost: $1.50/1k jobs (DIAMOND tier) → $5/1k (FREE tier) — 5–12× more expensive than Mode 1 Use when: the client needs recruiter contact data alongside the jobs (to outreach the recruiter, not the company), or AI-enriched job classifications (seniority extraction, skill tagging, normalised role taxonomy)

Returns everything Mode 1 returns, plus: recruiter LinkedIn URL, recruiter email, AI-classified seniority, AI-extracted skills, normalised role family.

Signal-strength matrix — voice-locked

Output bands:

• STRONG (signal_strength ≥ 7): worth a Tier-1 ABM placement, immediate routing
• MODERATE (4–7): Tier-2 / nurture-track candidate
• WEAK (1–4): note in /niche-signal-discovery output but don't prioritise routing
• NONE (0): no relevant hiring activity in window

Credit gate — voice-locked

/jobs-signal runs through .claude/rules/apify-credits.md's gating rules. Before any actor call:

1. Estimate cost: (expected_job_count × $/1k) / 1000 + actor_start_fee
2. Show estimate to user
3. Wait for confirmation if estimate >$5

Cost cheat sheet + invocation patterns in the premium reference.

Inputs

Process

Five-phase flow: Input validation → Cost estimation + credit gate → Actor call (Mode 1 or 2) → Signal extraction (job_count, role_types, most_recent_posting, target_role_match_count, signal_strength) → Output. Step-by-step + actor invocation patterns + cost cheat sheet in the premium reference.

Anti-Hallucination Guardrails

1. Never invent jobs. If the actor returns 12 jobs at Stripe, report 12, not "approximately 15."
2. Don't infer hiring intent from job descriptions beyond literal text. "Looking for someone with revenue-tooling experience" is signal. "Familiar with revenue tools" is not.
3. Date validation. If most_recent_posting is older than the date filter, flag — likely an actor bug, not a stale post.
4. Empty result handling. Zero jobs returned ≠ failure. It means no active hiring in the window. Report explicitly.
5. Recruiter data (Mode 2). If the actor returns a recruiter LinkedIn URL but no email, do not fabricate the email — pass to /deepline-enrich for waterfall lookup.

Quality

Pre-delivery checks cover coverage (date filter set, empty-result reporting), quality (counts match actor exactly, recruiter emails marked [UNAVAILABLE] when not returned, signal-strength applied per matrix), and cost discipline (Apify gate, Mode 2 only when justified). Worked example (ClientCo 100-firm sweep at $1.60) + anti-examples (Mode 2 by default, all-jobs-ever, fabricated recruiter emails, soft-language inference) + quality gate in the premium reference.

Chain Patterns

In the engagement workflow: slots into the AEO visibility loop as an optional context-input (a sudden hiring spike on a target account is a signal that AEO content for that account's space is timely) and into the sales pipeline as a discovery-prep enrichment.

Relationship to Other Skills

Upstream (consumes)

Downstream (feeds into)

Credit and Tool Reference

Final ship gate

Run /premortem --output before ship. See /premortem skill for the 5 execution domains (will-it-resonate / will-it-convert / will-it-stay-on-brand / will-stakeholder-push-back / will-it-degrade-over-time) and output template.

Trivial-case escape: ## Premortem\nNo failure modes — trivial change satisfies the contract for genuinely trivial outputs.