AI Tailoring Workflows

Once you're familiar with tags and views, an AI agent can handle the entire tailoring workflow for each job application.

Tailor Your Resume

Give the agent your resume.md and a job posting URL, and it handles tailoring, formatting, and verification.

The Agent's Workflow

1. Read the job description (URL or pasted text)
2. Map each requirement to your existing bullets: covered, weak, or missing
3. Decide what's durable vs ephemeral (see edit vs query)
4. Edit resume.md for durable improvements (new bullets, better phrasing, new tags)
5. Compose a view or CLI vars for ephemeral, per-JD adjustments (keywords, section order)
6. Render and show you the result

resumx resume.md --for stripe-swe -o out/stripe.pdf

Why you don't need to worry about layout

With pages: 1, Resumx automatically adjusts spacing and font size after every edit. The agent can add, remove, or rewrite bullets freely, the content will always fill exactly one page.

Example Prompt

Read resume.md and fetch <URL>. Identify the must-have requirements
from the job description, map each to my existing bullets, then
propose targeted edits. For keyword alignment and section ordering,
create a view instead of rewriting bullets. Keep facts truthful.

Replace <URL> with the job posting link. Adjust the instructions to match your workflow.

Edit vs Query

Not every change belongs in the resume file. The agent skill encodes a simple heuristic: will this change make the next 10 applications better, or just this one?

Situation	Action	Why
You shipped a new project	Edit resume.md	Every future application benefits
A JD emphasizes "stream processing"	View vars / CLI	Only this application cares
A bullet undersells its impact	Edit resume.md	Better phrasing helps everywhere
A role wants skills before experience	View pin	Other roles want the original order
You learned Rust	Edit resume.md	Permanent addition to your skillset
A JD says "CI/CD" and you wrote "deployment automation"	Judgment call	Could go either way

Durable changes grow your content library, they compound across every future application. Ephemeral changes live in views or CLI flags and don't clutter the source.

The mental model

Your resume.md is your career, and it grows as your career does. Each application is a different view on it. The agent's default mode is composing the right view (query), not rewriting the data (edit). It edits when the improvement is permanent, it queries when the adjustment is situational.

Agentic Workflows

Zero-File-Modification Rendering

For maximum speed and zero git diff pollution, the agent can render without touching any file:

resumx resume.md --for backend -v tagline="Stream Processing, Go, Kafka" --pin skills,work -o stripe.pdf

All inputs are explicit in the command. Nothing saved, nothing to undo.

Batch Applications with Views

Create .view.yaml files for each application:

# stripe-swe.view.yaml
stripe-swe:
  selects: [backend, distributed-systems]
  sections:
    pin: [skills, work]
  vars:
    tagline: 'Stream Processing, Event-Driven Architecture, Go, Kafka'

# vercel-fe.view.yaml
vercel-fe:
  selects: [frontend, ui]
  vars:
    tagline: 'React, UI Performance, Design Systems, Next.js'

The agent creates new .view.yaml files as you apply to new jobs. Each file is a complete, reproducible render job. Resumx discovers all *.view.yaml files automatically.

Progressive Commitment

The agent naturally escalates from ephemeral to persistent based on your usage:

1. One-off application: Pure CLI flags. Nothing saved.
2. Worth tracking: Agent creates a view. Reproducible, version-controlled.
3. Recurring role type: Agent adds a composed tag so content labels are reusable across multiple views.
4. Permanent improvement: Agent edits resume.md directly. Benefits all future renders.