job-search-powerhouse-skill

Job Search Powerhouse

A reusable Codex skill for running a job search like a senior recruiter, senior career counselor, and ethical interview coach.

It is built for one practical goal: help a candidate spend less effort on weak applications and more effort on roles that can realistically turn into interviews.

What It Does

job-search-powerhouse gives Codex a repeatable workflow for:

• resume intake and candidate profile setup
• recruiter-grade role fit checks
• ATS and human-readable resume tailoring
• market-intelligence scoring for hiring urgency and competition
• direct employer/ATS application workflows
• safe stopping rules for unsupported or sensitive fields
• email/status/rejection tracking when the user grants access
• outcome learning from callbacks and rejections
• OSINT interview preparation for positive-signal roles
• ethical assessment prep for coding, technical, support, and culture screens

The skill is intentionally generic. It does not include any specific user's resume, contact info, clearance, veteran status, or application defaults. New users must provide their own resume and intake answers.

Decision Framework

The skill uses a combined review stack before recommending or submitting an application:

• LENSSTACK: recruiter, hiring manager, ATS, compensation, competition, risk, and interview-prep lenses
• SWOT: strengths, weaknesses, opportunities, threats
• OODA: observe, orient, decide, act
• COMPARE: posting vs exact resume packet
• REDTEAM: try to reject the candidate before the recruiter does
• X10THINK: ask what would improve interview odds by 10x
• SIGNALVSNOISE: separate direct/fresh/niche signals from repost noise
• OPTIMIZE: prioritize interviews and offers over raw application count

Why It Stands Out

Most job-search automation optimizes for volume. This skill optimizes for recruiter believability.

It is designed to avoid:

• generic resume blasting
• senior-role overselling
• coursework pretending to be production experience
• stale or flooded job boards
• unsupported salary, clearance, schedule, or legal answers
• shallow "keyword match" decisions
• guessing interview questions without evidence

It is designed to create:

• fewer but stronger applications
• role-specific packets
• visible senior recruiter and senior career counselor gap reviews
• exact blocker ledgers
• rejection lessons that change future behavior
• callback analysis that identifies what worked
• interview prep grounded in public evidence and the actual posting

It also treats ATS anti-spam failures as first-class operational blockers. If a fully supported application is rejected by the platform itself, the skill should save screenshots, preserve the exact answers/files used, and hand off a clean manual retry path instead of pretending the packet was unfinished.

Install Locally

Copy the skill folder into your Codex skills directory:

cp -R skills/job-search-powerhouse ~/.codex/skills/

Restart Codex, then ask:

Use $job-search-powerhouse to intake my resume and run a recruiter-grade remote technical support job search.

First Run

The skill should ask for the user's resume and intake details before applying anywhere.

Start with:

Use $job-search-powerhouse. Here is my resume. Build my profile, find strong-fit jobs, and stop before submitting anything until I approve.

The intake template lives at:

skills/job-search-powerhouse/assets/templates/intake-questionnaire.md

Repository Layout

skills/job-search-powerhouse/
├── SKILL.md
├── agents/openai.yaml
├── assets/templates/
│   ├── application-tracker.md
│   ├── intake-questionnaire.md
│   ├── interview-osint-report.md
│   ├── sample-intake.json
│   └── sample-posting.txt
├── references/
│   ├── ethical-assessment-prep.md
│   └── recruiter-proof-rubric.md
└── scripts/
    ├── init_user_profile.py
    └── score_role.py

Validation

Run the full quality harness:

python3 tests/run_quality_checks.py

Expected result:

PASS test_skill_validation
PASS test_no_personal_leakage
PASS test_sample_json_and_profile_generation
PASS test_role_scoring_expected_behavior
PASS test_next_user_dry_run_fixtures
PASS test_portable_install_copy
ALL QUALITY CHECKS PASSED

The publish-readiness proof is documented in:

PUBLISH_READINESS_REPORT.md

Fresh-clone human-style testing is documented in:

HUMAN_STYLE_TEST_REPORT.md

Example Helper Usage

Create a generic profile from intake JSON:

python3 skills/job-search-powerhouse/scripts/init_user_profile.py \
  skills/job-search-powerhouse/assets/templates/sample-intake.json \
  -o profiles/example-profile.md

Score a posting text file:

python3 skills/job-search-powerhouse/scripts/score_role.py \
  skills/job-search-powerhouse/assets/templates/sample-posting.txt

Safety Boundaries

The skill does not guarantee interviews or offers.

It should not:

• invent credentials or employment facts
• submit applications without permission
• read unrelated email
• answer legal, background, demographic, or self-ID fields without support
• bypass CAPTCHA or account controls
• provide stolen live assessment answers
• fake psych/personality traits

It should:

• ask for missing facts
• stop on unsupported fields
• preserve user privacy
• prepare ethically for interviews and assessments
• turn outcomes into improved targeting and resume strategy

Resume Portfolio Use

Safe resume bullet:

Built and validated a reusable Codex job-search skill that performs resume intake, recruiter-grade role targeting, market-intelligence scoring, ATS resume tailoring, ethical application handling, rejection analysis, outcome learning, and OSINT interview preparation.

Project description:

Created a portable AI-agent workflow with reusable templates, validation scripts, privacy checks, and dry-run fixtures to help candidates identify strong-fit roles, avoid overselling, tailor recruiter-ready packets, track outcomes, and prepare for interviews using evidence-driven OSINT.