β³
Loading job data...
π
No jobs match your filters
Try adjusting the filters or picking a different stage
Principal Β· Staff Β· Group Product Manager
Status changes are saved in this browser's local storage by default β they won't show up on your phone or another browser.
Paste a GitHub token here to also write status changes straight back to jobs_history.json in the repo, so every device sees the same status.
Use a fine-grained personal access token scoped to only this repository (dibajaj/Job-search-agent), with
Contents: Read and write permission and nothing else. The token is stored only in this browser's local storage and sent only to api.github.com.
Paste a LinkedIn job URL to queue it for scoring and resume generation.
Then run /add-job in Claude Code to process it.
No GitHub token configured β copy this command and run it in Claude Code:
β
Your reason helps calibrate scoring in future runs β be specific.
β³
Loading job data...
π
No jobs match your filters
Try adjusting the filters or picking a different stage
π
No run data yet
This fills in after /run-job-search has logged at least one run
Success Rate
β
Runs Shown
β
Avg Duration
β
Avg Batches/Run
β
Run History
Bar = seconds per resume generated (green = success, red = failed β hover a red bar for the failure reason). Line = tokens used, when available.
Failure Breakdown
No failures in this window π
Yield Funnel β Average Per Run Over Window
Task Time Breakdown
Total: β
Most recent run with task-level timing data
Why Jobs Are Dropped Below Threshold β Over Window
| Reason | Frequency |
|---|
No drop-reason data yet β available from the next run.
Observability for Claude's own scoring/resume-drafting judgment, as opposed to the deterministic run health above.
Applied vs. Deleted Profile
Deleted is the real negative signal (not "never applied", which is ambiguous) β the empirical base the preference check-in (/check-preferences) draws from.
π¨ Applied / Interviewing (0)
Avg score: β
Score range: β
ποΈ Deleted (0)
Avg score: β
Score range: β
Resume Eval Results β 1-in-10 Sampled, Opus-Judged
From eval_log.jsonl, written inline during /run-job-search Step 4c. Score out of 100: Grounding /50 Β· Job Relevance /30 Β· Tailoring /10 Β· Format /10.
| Company | Sampled At | Grounding | Tailoring | Format | Score | Opus Notes |
|---|
No resume evals logged yet β the first lands on the 10th resume drafted in a run.
Preference Check-in History
From preference_checkins.jsonl, written by /check-preferences.
| Date | Pattern Found | Answer | Config Change |
|---|
No check-ins run yet β run /check-preferences once there's enough Applied/Deleted data.
β Gap Warning
Job ID