jobra

Jobra bridges the gap between raw resumes and high-signal job discovery. It moves beyond simple keyword matching by enforcing a strict evidence-linked workflow. This skill extracts data from local resumes to create verified candidate profiles, manages separate search preferences, and audits live job postings. It prioritizes data integrity, ensuring every claim in a profile or match rationale is backed by a specific excerpt from the source documents.

What it does

• Profile extraction transforms local resumes into evidence-linked JSON drafts for human verification.
• Evidence-backed matching explains candidate-opportunity alignment by citing specific document excerpts for every requirement.
• Opportunity discovery researches live job postings and creates validated, deduplicated opportunity snapshots.
• Result auditing generates private shortlists that highlight career alignment, identified gaps, and data freshness.
• Conflict separation keeps candidate profiles, search preferences, and job data in distinct, immutable schemas to prevent context drift.

How it works

1. Setup phase triggered by the setup argument to parse a local resume and generate a verifiable profile draft.
2. Profile verification via the profile command to audit evidence links and confirm the data contract.
3. Discovery runs using discover to research postings and save immutable snapshots to the local .jobra/ directory.
4. Matching and results invoked through match or results to evaluate fit against specific opportunity IDs or titles.

Frameworks & tools

The skill utilizes Python scripts for deterministic validation and schema enforcement (jsonschema). It manages local state via a structured .jobra/ directory and integrates with Codex for evidence-linked reasoning and web-based research.

Why this beats prompting it yourself

Generic prompts often hallucinate qualifications or miss subtle requirement gaps. Jobra enforces a strict evidence-sufficiency policy and deterministic validation, ensuring that match rationales are grounded in local files rather than model training data.

Use cases

• Build a verified, machine-readable candidate profile from a legacy PDF resume.
• Audit a batch of job postings against specific technical constraints and preferences.
• Generate a gap analysis report for a specific job opportunity to prepare for interviews.
• Maintain a private, deduplicated history of researched job opportunities and fit assessments.

Known limitations

Does not support automatic job applications. Requires manual verification of all profile drafts. Research and live discovery rely on external reasoning tools rather than a local search backend.