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.

Your profile is the foundation of every application AI Job Search produces. The richer and more specific the content in your profile files, the more precisely the system can tailor your CV, evaluate job fit, and write cover letters that speak directly to a specific role. This page explains what each file contains, how they are used during /apply and /scrape, and which ones /setup populates automatically versus which ones you maintain by hand.

File overview

FileUsed byPopulated byEdit manually?
CLAUDE.md/apply, /scrape, all commands/setupYes, for profile updates
01-candidate-profile.md/apply Steps 1-4/setupYes
02-behavioral-profile.md/apply Step 3 (reviewer)/setupYes
03-writing-style.md/apply Step 2Shipped, extended by /setupYes
04-job-evaluation.md/apply Step 1/setupYes
05-cv-templates.md/apply Step 2/setupYes
06-cover-letter-templates.md/apply Step 2ShippedRarely
07-interview-prep.mdOn-demand interview prep/setup stubs; you completeYes
search-queries.md/scrape/setupYes

CLAUDE.md: the main profile

CLAUDE.md is the file Claude reads in every session as the primary profile for the workspace. It contains your identity, career goals, and the framework’s workflow rules and verification checklist. When you run /setup, every [PLACEHOLDER] token in CLAUDE.md is replaced with your actual information. The structure, workflow rules, and verification checklist are preserved unchanged. What gets filled in:
  • Identity: name, location, languages, employment status, LinkedIn headline
  • Education: degrees in reverse chronological order with thesis topics where applicable
  • Professional Experience: roles with responsibilities and key achievements
  • Technical Skills: primary skills, secondary skills, domain expertise, tools
  • Behavioral Profile: assessment results or synthesized self-assessment
  • What Excites You: the tasks and environments that energize you professionally
  • Target Sectors: industries and example companies you are pursuing
  • Deal-breakers: hard constraints that disqualify a role automatically
The Verification Checklist section in CLAUDE.md is not profile data — it is the checklist the drafter runs at Step 6 of every /apply run. It covers factual accuracy, targeting, consistency, quality, compiled PDF layout, and ATS text-layer checks. You should not edit this section unless you want to change the verification rules for your workflow.

The seven skill files

The seven files under .claude/skills/job-application-assistant/ hold the structured data and guidance that power the /apply pipeline. They are read selectively at each step to keep token use efficient.

01-candidate-profile.md

Read at Step 1 (fit evaluation) and kept in context through Step 4. Contains your structured professional data in a format optimized for the application workflow:
  • Identity block: name, contact details, location, languages, constraints
  • Education table: degree, period, institution, key topics
  • Professional Experience: one section per role, with responsibility and achievement bullets
  • Independent Projects: open source, freelance, or personal projects outside employment
  • Technical Skills: programming languages with proficiency levels and frameworks, domain expertise, software and tools
  • Publications, Awards, References
The level of detail you put here directly shapes the quality of CV tailoring. Generic bullets produce generic applications; specific achievements with tools, data volumes, and measurable outcomes give the drafter material to work with.

02-behavioral-profile.md

Read by the reviewer agent at Step 3. Contains your behavioral assessment results (PI, DISC, Myers-Briggs, StrengthsFinder, or a self-assessment) structured for use in application review:
  • Core Behavioral Drives: scored drives with descriptions
  • Strongest Behaviors: your most prominent working style traits
  • How You Work Best: environment and collaboration preferences
  • Growth Areas: framed constructively for interview use
  • Mapping to Job Posting Language: keywords that signal strong behavioral fit versus potential friction
  • Management Style Preferences: what works and what does not
The reviewer uses this file specifically to check whether the cover letter’s voice matches your natural register. A profile that scores high on “Collaborator” traits should not be given a solo-hero tone in the cover letter; a “Persuader” profile should not be given over-hedged, apologetic phrasing.

03-writing-style.md

Read at Step 2 (drafting) and by the reviewer at Step 3. Contains the rules every cover letter and CV must follow:
  • Critical rules: no em-dashes, no clichés, no generic buzzwords without backing, no apologetic language, no unverified company claims
  • The interview backtrack test: a check for whether a reframed bullet could be explained honestly in an interview without backtracking
  • Tone guidance: warm but direct, conversational professional, first person active voice
  • Cover letter structure: opening, body, company-specific paragraph, closing
  • Language guidance for different role types: technical/ML, domain-specific, consulting, leadership
  • Multi-language rules: default to the posting’s language; the letter should feel natural, not translated
This file ships with content and does not need to be populated by /setup. If you want to add patterns observed in your past applications (which /setup Path A does automatically when it reads your applications/ folder), those go under a ## Patterns Observed in Past Applications section.

04-job-evaluation.md

Read at Step 1. Contains the scoring framework the drafter uses to evaluate every job posting:
  • Five scoring dimensions with rubrics: Technical Skills Match (0-100), Experience Match (0-100), Behavioral/Culture Fit (0-100), Location & Logistics (Pass/Fail), Career Alignment (0-100)
  • Your strong/moderate/weak match areas: populated by /setup with your actual skills
  • Career goals and motivation filters: what tasks energize you, what drains you, what life-situation constraints apply
  • Weighting: Technical Skills 30%, Experience Match 25%, Behavioral Fit 15%, Career Alignment 30%
  • Thresholds: Strong Fit 75+, Good Fit 60-74, Moderate Fit 45-59, Weak Fit 30-44, Poor Fit below 30
  • Pre-application call guidance: when and how to call the contact person before submitting
The skill match areas section is the most important part to keep current. As you gain new skills or shift priorities, update the strong/moderate/weak lists to keep evaluations accurate.

05-cv-templates.md

Read at Step 2 (drafting). Contains the full LaTeX structure and tailoring rules for the CV:
  • Template specification: moderncv banking style, lualatex compile command, expected output
  • Document structure: preamble with color overrides, required \renewcommand lines for correct rendering on modern MiKTeX
  • Section-by-section tailoring rules: profile statement templates per role type, skills section reordering, experience bullet guidance, publication selection, reference formatting
  • Page budget: hard 2-page limit with per-section content budgets
  • Relevance-weighted cutting: the algorithm for shrinking a CV when it overflows (score by posting relevance, uniqueness, cover letter dependency; cut lowest scorer first)
  • Compile-and-inspect loop: step-by-step instructions for layout verification and common fixes
  • ATS parseability guidance: what to check in the pdftotext extraction
/setup adds role-specific profile statement templates here based on your background. If you register a custom LaTeX template with /add-template, its instructions are also stored here.

06-cover-letter-templates.md

Read at Step 2. Contains the full cover letter template and compile instructions:
  • Template specification: cover.cls, xelatex compile command, font directory
  • Full document structure: \namesection{}, \lettercontent{}, \closing{}, \signature{}
  • Known template pitfall: the \lettercontent{} + \begin{itemize} interaction — placing an itemize environment inside \lettercontent{} breaks the compile because the macro appends \\ after \end{itemize}. The correct pattern closes \lettercontent{} before the list and wraps the list in a \fontspec block to preserve the Raleway-Medium font.
  • Tailoring guidelines: salutation rules, word budget (250-300 words of body text), bullet list style, non-English cover letter rules
  • Finalization checklist: no em-dashes, forward-looking framing, company name and role correct, language matches posting
This file ships complete and rarely needs manual editing. If you want to record patterns from past cover letters, /setup Path A can add observations here from your applications/ folder.

07-interview-prep.md

Used on-demand when you ask for interview preparation for a specific role. Contains:
  • STAR format guidance: structure (Situation, Task, Action, Result) and timing advice
  • Ready-made STAR examples: populated by /setup from your actual experience. Path A (documents folder) creates stubs with a source tag and asks you to complete the S/T/A/R fields manually; Paths B and C generate full examples.
  • Common tough questions: pre-written answers for standard interview challenges, personalized to your situation
  • Questions to ask interviewers: grouped by role, team, tech and growth, and culture
  • Phone and video interview tips
  • Follow-up etiquette and thank-you note guidance
  • Roleplay guidelines: how to request interview practice from Claude
If you ran /setup via Path A (documents folder), check 07-interview-prep.md for ## STAR Candidates (Complete Manually) stubs. These are achievement skeletons extracted from your CV and reference letters. Filling in the S/T/A/R fields before your next interview session gives Claude concrete material to work with during roleplay practice.

search-queries.md

Located at .claude/skills/job-scraper/search-queries.md. This file controls everything /scrape searches for. What it contains:
  • Query categories (Priority 1-4): Site-scoped WebSearch queries grouped by career direction. Priority 1 is your strongest target role type; Priority 4 is a broad fallback.
  • Location filter: Acceptable commute areas defined as tiers (ideal, acceptable, borderline, too far). Jobs outside this range are skipped automatically.
  • Date filter: Only postings from the last 14 days are included; expired deadlines are skipped.
  • Adapting for focus areas: When you pass a focus-area argument to /scrape, queries from the matching category are prioritized and additional custom queries are generated.
Populated by: /setup Section 9 (or the equivalent follow-up questions in Path A). It replaces placeholder tokens like [YOUR_PRIMARY_JOB_TITLE] and [YOUR_CITY] with your actual role titles, skills, and location. How to update: Edit the file directly as your search priorities change, or run /setup --section search for a guided update interview that also suggests role types you may not have considered.

How the files work together

During a /apply run, the files are read in a specific order designed to avoid redundant reads:
Step 1 (Fit Evaluation):
  → reads 04-job-evaluation.md + 01-candidate-profile.md
  → both stay in context for all remaining steps

Step 2 (Drafting):
  → reads 03-writing-style.md + 05-cv-templates.md + 06-cover-letter-templates.md
  → reads one existing cv/*.tex and one cover_letters/*.tex as structural reference

Step 3 (Reviewer Agent — fresh context):
  → receives CV and cover letter drafts inline (no file read)
  → reads 01-candidate-profile.md + 02-behavioral-profile.md + 03-writing-style.md + 04-job-evaluation.md

Step 4 (Revision):
  → no new file reads; uses context from Steps 1 and 2
  → verifies company claims via WebFetch/WebSearch before including

Step 5 (Compile & Inspect):
  → reads 05-cv-templates.md and 06-cover-letter-templates.md for fix guidance if needed

Step 6 (Verification):
  → reads CLAUDE.md verification checklist
  → re-reads both output files once to confirm final state on disk
During a /scrape run, the skill reads search-queries.md for query configuration and does a quick fit check against your profile, but does not read the full skill files. Those are only loaded when you select a job from the results and the full /apply workflow begins.

Which files /setup populates

Populated automatically by /setup

CLAUDE.md, 01-candidate-profile.md, 02-behavioral-profile.md, 04-job-evaluation.md, 05-cv-templates.md, 07-interview-prep.md, cv/main_example.tex, search-queries.md

Ships with content, extended by /setup

03-writing-style.md (ships complete; Path A extends with observed patterns), 06-cover-letter-templates.md (ships complete; Path A may add structural patterns from past applications)
If you prefer editing files directly instead of using /setup, all files are plain Markdown and can be updated in any text editor. The [PLACEHOLDER] structure makes it straightforward to fill in your information section by section.

Build docs developers (and LLMs) love