Staleness Detector skill

Detects stale, dead, and low-quality job postings early.

Script

staleness_detector.py

Analyzes posting metadata and detects staleness/quality issues. Accepts posting metadata dict, returns staleness score (0-10), status (FRESH/ACCEPTABLE/AGING/STALE), red flags, and recommendation.

Call detect_staleness(posting_metadata) to assess posting quality and viability.

Staleness Score:

• •8+: FRESH (green light)
• •6-8: ACCEPTABLE (yellow light, check red flags)
• •4-6: AGING (orange light, human review)
• •<4: STALE (red flag, likely dead)

Red Flags Detected:

• •Posting age > 30 days or > 60 days
• •Applicant volume > 200 or > 500
• •Contradictory technologies
• •Unrealistic experience requirements
• •Vague JD (too short)
• •Missing technologies or experience years
• •Missing salary or location
• •Duplicate posting pattern
• •Reposting within 7 days of same role

Output includes:

• •Staleness score and status
• •Recommendation (green/yellow/orange/red light)
• •Age analysis (days old, category)
• •Volume analysis (applicant count, category)
• •Complete red flag list
• •Boolean flags: is_dead, needs_human_review

Rules

• •Never submit without checking staleness first.
• •Posting > 30 days old gets escalated to human review.
• •Posting > 500 applicants is likely dead or viral posting.
• •Multiple red flags trigger escalation regardless of age.
• •Staleness_detector runs first in Jay's workflow before deep analysis.