How the ghost-likelihood score works
The score is the share of 114 weighted risk points a listing triggers, normalized to 0-100. It is probabilistic — a density of risk signals, never proof, and never an accusation against a named company.
Step 1: Paste the listing
Copy the full job description (or its URL) into the checker.
Step 2: The rule engine scores it
11 deterministic detectors test the text against known ghost-job patterns — no AI, no randomness, same input always gives the same score.
Step 3: Read the evidence
Every triggered flag shows the exact matched phrase and a plain-English why, linked to a full explainer with sources.
Step 4: Verify the hidden signals
Posting age and repost history are often hidden by boards — the result tells you what to check manually.
The weights
| Signal | Points | Share |
|---|---|---|
| No salary range listed | 18 | 15.8% |
| Posting older than 30 days | 16 | 14.0% |
| Reposted again and again | 14 | 12.3% |
| Boilerplate, buzzword-heavy description | 12 | 10.5% |
| Vague or stacked job title | 10 | 8.8% |
| 'Always hiring' / talent-pipeline language | 10 | 8.8% |
| No named hiring manager or contact | 8 | 7.0% |
| No concrete details anywhere | 8 | 7.0% |
| Urgent hiring, zero specifics | 6 | 5.3% |
| Implausibly wide salary range | 6 | 5.3% |
| Hidden or masked employer | 6 | 5.3% |
Score bands
- 0-24 — low: few known patterns; worth your time.
- 25-54 — moderate: verify the flagged items first.
- 55-100 — high: matches many ghost patterns; verify independently before investing 45 minutes.
Key takeaway: the engine is rule-based and open. The same listing always scores the same;
the weights and data are published at /api/v1/flags and /api/v1/boards.
Check a listing now: the free checker tests this flag (and the other ten) against any pasted listing — with the matched evidence. No signup.