Contenu du pack

Un fichier SKILL.md (le guide pratique) + un modèle de départ. Aperçu :

AI Job Auto Applier

The facts this skill runs on (verified 2026-11-02)

• The pipeline is five stages, and submission is never automatic. Profile → match → tailor → approval queue → submit. The assistant does the research and writing; you approve each application individually. "Auto" covers the prep work — the send button is always yours.
• Tailoring is honest reframing, never fabrication. The assistant rewords your real experience to match a posting's language. It never invents jobs, titles, dates, degrees, or skills you don't have. A tailored resume that invents history isn't tailored — it's fraud, and this skill refuses to write it.
• Job boards have rules; the skill respects them. Many boards prohibit automated submissions or require applications through their own flow. The skill checks the board's terms for each posting and tells you when a posting must be applied to manually — it doesn't route around ToS.
• Quality beats volume. Ten tailored applications outperform a hundred generic ones. The skill is built around a short list of strong matches, not spray-and-pray.

When to use

• You're job hunting and the per-posting customization is eating your evenings — the assistant drafts, you approve.
• You want every application tailored to the posting without rewriting your resume from scratch each time.
• You need a tracked queue: what you applied to, when, with which resume version, and what came back.
• You're returning to the market and want your materials reviewed against current postings.

Safety rules (non-negotiable)

• You approve every single submission. The approval queue shows the tailored resume, the cover letter, and the posting side by side. No batch-send, no "apply to all."
• Nothing fabricated, ever. No invented employers, titles, dates, degrees, certifications, or skills. If the posting asks for something you don't have, the skill says so instead of inventing it.
• Respect each board's terms. If a board prohibits automated application, the skill prepares the materials and you submit manually through their flow. No CAPTCHA-solving, no ToS circumvention.
• Your data stays yours. Resumes, contact details, and application history live in your own files. Nothing is uploaded anywhere except the applications you personally approve and submit.

Workflow

1. Profile build (input: your background → output: master profile)
   One thorough intake: work history, skills, education, what you're looking for (roles, locations, remote, salary band), and deal-breakers. The assistant writes a master profile and a base resume. You review it once — everything downstream builds on this, so accuracy matters more than polish here.
2. Job matching (input: master profile + search criteria → output: ranked shortlist)
   You point the assistant at postings (links, or a board search you ran). It scores each against your profile: skill overlap, seniority fit, location/remote match, and red flags (vague description, "urgent hire" + below-market pay, mismatched title). Output is a ranked shortlist with one-line reasons — typically 5–10, not 50.
3. Tailoring (input: shortlist → output: tailored resume + cover letter per posting)
   Per posting, the assistant produces: a tailored resume (reordered bullets, posting keywords reflected honestly, irrelevant roles condensed) and a short cover letter (why this role, one concrete proof point). Each pair notes exactly what changed versus the base resume — so you can see the tailoring, not just trust it.
4. Approval queue (input: tailored pairs → output: your decision per posting)
   You see each posting next to its tailored materials and the match rationale. Per posting: approve (moves to submit), edit (your notes → revised), or skip (logged with reason). Nothing leaves the queue without your explicit approve.
5. Submit (input: approved applications → output: submission log)
   You submit through the board's normal flow (or the assistant walks you through it click by click). Each submission is logged: posting, date, resume version, cover letter version, and confirmation. Follow-ups and responses get tracked against the same log.

Aperçu du modèle

AI Job Auto Applier — Job Report

One report per application. Fill every field — "n/a" is an answer, a blank is not.

Master profile

• Profile version / date reviewed:

Match

• Posting (title / company / link):
• Match rationale:
• Red flags: none / notes

Tailoring

• Resume version:
• Changes vs. base resume:
• Cover letter version:
• Honesty check: no invented employers, titles, dates, degrees, or skills

Approval

• [ ] Approved — submit
• [ ] Edit — notes:
• [ ] Skipped — reason:
• Decided by / date:

Submission

• Submitted at:
• Board / method (manual per ToS where required):
• Confirmation ref:

Le pack de compétence lui-même (SKILL.md + modèle) est rédigé en anglais.

Acheter le pack complet — $4.99 Pro — les 531 compétences

What does AI job auto apply actually automate?

The prep work: building your master profile, scoring postings against it, tailoring each resume and cover letter to the posting's language, and tracking everything in a queue. The submission itself is never automatic — you review each tailored pair and approve it individually.

Will it invent experience to match a posting?

Never. Tailoring means reframing your real experience in the posting's language — reordered bullets, honest keyword reflection. If a posting asks for something you don't have, the skill flags it instead of inventing it. Fabricated employers, titles, or skills are a hard refusal.

Does it work with every job board?

It prepares materials for any posting, but respects each board's terms: where automated submission is prohibited, the skill gets everything ready and you submit manually through the board's own flow. No CAPTCHA-solving, no terms circumvention.