Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/SudharakaA/ai-job-search/llms.txt

Use this file to discover all available pages before exploring further.

Running /setup is the fastest way to populate your profile from scratch, but once your profile is in place you rarely need to re-run the full interview. Most updates, including adding a new job, refining your career goals, tweaking evaluation criteria, or adjusting search queries, are faster to make by editing the relevant file directly. Understanding what each file controls helps you make targeted changes without disturbing the parts you want to keep.

Which files to edit manually

FileWhat to change
CLAUDE.mdYour full profile: name, education, experience, skills, goals, target sectors, deal-breakers
.claude/skills/job-application-assistant/01-candidate-profile.mdStructured version of your CV data: education entries, job entries, skills lists
.claude/skills/job-application-assistant/02-behavioral-profile.mdYour behavioral assessment or self-assessment (PI, DISC, Myers-Briggs, or your own framing)
.claude/skills/job-application-assistant/04-job-evaluation.mdSkill match areas, career goals, and motivation filters used when scoring job fit
.claude/skills/job-application-assistant/05-cv-templates.mdProfile statement templates for different role types
.claude/skills/job-application-assistant/07-interview-prep.mdYour STAR examples drawn from actual experience
.claude/skills/job-scraper/search-queries.mdJob search queries for /scrape: roles, skills, locations, portals

The role of CLAUDE.md

CLAUDE.md is the main candidate profile and the first file Claude reads in every session. It contains your name, location, languages, employment status, education, professional experience, technical skills, certifications, publications, awards, behavioral profile, motivations, target sectors, and deal-breakers. Everything /apply and /scrape do is grounded in this file. /setup populates CLAUDE.md automatically, replacing [PLACEHOLDER] tokens with your actual information. After that, you own it. Treat it as a living document: update it when you change jobs, complete a significant project, earn a certification, or clarify your career direction.

Framework rules vs. candidate data

Not all profile files are pure candidate data. Two files contain framework rules that shape how the system evaluates and writes applications: 03-writing-style.md defines the tone, structure, and stylistic do’s and don’ts for all generated text. It ships with sensible defaults. You can adjust it to match your voice, but read it before making changes so you understand what each rule does. 04-job-evaluation.md contains both your personal skill match areas and career goals (candidate data you should update) and the scoring framework and evaluation criteria (framework rules that govern how fit is assessed). When editing this file, update the sections that describe your goals and priorities freely. Be careful about modifying the evaluation criteria themselves unless you have a specific reason to change how fit is scored.
Editing 03-writing-style.md or 04-job-evaluation.md changes how the system behaves for all future applications, not just the next one. Read the file in full before making changes so you know what you are adjusting.

Updating search queries

Search queries live in search-queries.md and control what /scrape searches for: which roles, which skills, which locations, and which portals. As your job search evolves (different roles come into focus, you add a new skill area, or you expand to new cities) the queries need to evolve with it. You have two ways to update them:
Run the search configuration section of /setup without touching the rest of your profile:
/setup --section search
This re-runs the search configuration interview: which roles to target, which skills to search for, which locations, and which portals. It also suggests role types you may not have considered based on your full profile. Use this when you want a guided update that cross-references your profile against new query ideas.

Profile depth and output quality

The quality of generated CVs, cover letters, and fit evaluations scales directly with the depth of your profile. Thin profiles produce generic output; detailed profiles enable genuinely tailored applications.
Do not just list job titles. For each position in 01-candidate-profile.md and CLAUDE.md, describe what you actually did: specific projects, tools used, responsibilities, and measurable achievements. The more material you provide, the more precisely the system can reframe your experience for different roles.A single detailed entry like “Built ML pipelines for customer churn prediction in Python using scikit-learn, reducing churn by 12% over six months” gives the system far more to work with than a job title plus a list of technologies.
Rather than listing Python or project management as bare skills, describe how and where you applied them. Context turns a skill list into evidence: “Led cross-functional teams of 4–8 people across three product launches” is a different claim than “project management,” and the system can use it differently across different role types.
The 02-behavioral-profile.md and the motivational sections of CLAUDE.md (what energizes you, what you want more of, what drains you) directly shape how the system evaluates culture fit and which roles it surfaces as strong matches during /scrape. These sections are worth investing time in during initial setup and revisiting as your priorities evolve.

Career path discovery

The framework supports two distinct modes when evaluating job fit: Explicit targeting works when you know which roles or sectors you want. Set your target roles and sectors clearly in CLAUDE.md and 04-job-evaluation.md, and the system focuses fit evaluation and search queries on those targets. Latent opportunity discovery works when you want the system to surface roles you have not explicitly considered. By analyzing your full work history (not just job titles, but the actual work described in your profile), the system can identify transferable skills that map to unexpected industries, patterns in what you excelled at, and emerging roles that combine your domain expertise with new technology. This mode requires richer profile content: the more context you provide about what you actually did and what you enjoyed, the more the system has to work with when identifying non-obvious paths. To get the most from latent discovery, include in CLAUDE.md not just your experience but what energized you in each role, what you would want more of, and what kind of work you found draining. This context shapes both fit evaluation and which roles /scrape surfaces.

Build docs developers (and LLMs) love