agentleFS
alexmonegropalexmonegrop/job-hunt-os/.claude/skills/apply/SKILL.md

apply

Skill0 starsChanged 7 months agoMITView on GitHub

End-to-end application workflow: resume tailoring, cover letter drafting, form submission via Playwright (non-LinkedIn) or LinkedIn Easy Apply automation via browser_run_code shadow DOM pattern, database record creation. Screenshots before submission.

28 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use outreach, track and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • QuixiAISkill

    outreach

    Decide whether a message is worth someone's attention, then reach the user through the outbox, or people on a connected chat channel, on their behalf or your own initiative

    Also in 5 projects that use this

  • theYahiaSkill

    track

    Отследить посылку Почты России

    Also in 4 projects that use this

  • NVIDIASkill

    evaluate

    Scores existing Sports Intelligence AVLM inference outputs with avlm/evals: MCQ accuracy (run_mcq_eval.sh, INFERENCE_DIR), open-ended LLM-as-judge scores (needs CLIENT_API_KEY for the configured judge endpoint), tennis category breakdowns, run comparisons, chance-level sanity checks and triage of mcq_results.json or llm_judge_results.json. Use when the user asks to score, evaluate, benchmark or compare predictions.jsonl from inference runs, or to explain a suspicious accuracy. Not for running inference, training, quantization, data preparation, or third-party API model inference: never open, list or run anything under avlm/evals/inference_3rdparty, including when the user asks for it; it is excluded from the release.

    Also in 4 projects that use this

  • ninehillsSkill

    hunt

    Finds root cause before applying fixes for errors, crashes, regressions, failing tests, broken behavior, and screenshot-reported defects. Use when users report in any language errors, crashes, broken behavior, regressions, failing tests, screenshot evidence, or something that used to work and now fails. Not for code review or new features.

    Also in 3 projects that use this

  • ledqSkill

    tailor

    Tailor the resume to one job posting. Give it a URL, paste the posting (or refer to one pasted earlier in this chat), or point at a saved application folder; it resolves the JD and runs the tailor-pipeline workflow end to end.

    Also in 3 projects that use this

  • proficientlyjobsSkill

    tailor-resume

    Tailor your resume for a specific job posting

    Also in 4 projects that use this

The file

Show its rules and text

We split this skill into the rules it gives an agent

23 rules. 0 are common, 13 are distinctive, and 10 are very specific to this project.

  • 0

    Common

    Most of the 27 edited versions we read keep these.

  • 13

    Distinctive

    This author's own choices. Worth a look.

  • 10

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Check the Job Posting's adjusted score:

    line 41
  2. Very specific:

    If the Job Posting has no adjusted score, calculate it first using the scoring table in the job-search skill (which reads …

    line 47
  3. Distinctive:

    Check if that file exists for THIS EXACT job posting.

    line 55
  4. Distinctive:

    Use the closest matching generic template (Innovation / Delivery / Generic) — no full tailoring.

    line 60
  5. Distinctive:

    NEVER upload a resume tailored for Company A to an application for Company B.

    line 64
  6. Distinctive:

    NEVER substitute a "closest match" tailored resume.

    line 65
  7. Distinctive:

    Before uploading ANY resume, verify the file path contains the target company name OR is a generic template.

    line 66
  8. Distinctive:

    Read Job Posting record from the database.

    line 73
  9. Distinctive:

    Verify tailored resume exists (Step 0 gate).

    line 74
  10. Very specific:

    Create Application record (status: Submitted)

    line 114
  11. Very specific:

    Create Interaction record (type: Application Submitted)

    line 115
  12. Very specific:

    Update Job Posting status to Applied

    line 116
  13. Distinctive:

    Read Job Posting record — get the LinkedIn job ID.

    line 121
  14. Very specific:

    If browser_navigate times out on the snapshot, that's OK — the page is loaded.

    line 132
  15. Very specific:

    Use a SINGLE browser_run_code call to handle the entire flow:

    line 136
  16. Very specific:

    Create Application record (status: Submitted, method: Easy Apply).

    line 255
  17. Distinctive:

    DO NOT hardcode standard answers in this skill — they're per-user.

    line 270
  18. Distinctive:

    Should the user reach out to the contact first?

    line 278
  19. Very specific:

    Use browser_run_code to batch combobox selections efficiently.

    line 299
  20. Very specific:

    Always browser_snapshot before filling.

    line 308
  21. Very specific:

    Always browser_take_screenshot before and after submit.

    line 309
  22. Distinctive:

    ALWAYS check for warm-application opportunity: before applying.

    line 316
  23. Distinctive:

    ALWAYS create both Application + Interaction records: after submission.

    line 317

“Common” is measured: kept by at least 6 in 10 of the 27 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Apply

End-to-end job application workflow with Playwright automation for ALL platforms including LinkedIn Easy Apply.

What This Skill Does

  • Resume tailoring — calls the resume-tailor skill to create role-specific version
  • Cover letter drafting — 250-400 word cover letter with company-specific hook
  • Non-LinkedIn automation — Playwright form filling for Indeed, Glassdoor, company portals
  • LinkedIn Easy Apply automation — Playwright via browser_run_code with shadow-DOM pattern (user-authorized)
  • Screenshot verification — captures pre-submit screenshot
  • Database integration — creates Application + Interaction records
  • Cross-workstream — links to existing networking contacts when available

Prerequisites

Required:

  • Playwright MCP (load via ToolSearch("playwright browser") each session)
  • Tailored resume (generated via resume-tailor skill)
  • Job Posting record in the database
  • applications/resumes/data/{user-slug}/standard-answers.yaml
  • applications/resumes/data/{user-slug}/user-config.yaml
  • Multi-user: resolve paths using the active user's slug (JOB_HUNT_USER).

Quick Start

"Apply to the Senior Program Manager role at Acme Renewables (Job Posting record exists)"
"Apply to my top 5 scored job postings"

Workflow

Step 0: Adjusted-score check (MANDATORY — BEFORE EVERYTHING)

Check the Job Posting's adjusted score:

  1. Adjusted ≥ 7 → full protocol: tailored resume + cover letter + portal-first
  2. Adjusted 5-6 → longshot protocol: Easy Apply only, closest generic template, NO cover letter, NO portal hunt. Tag "LONGSHOT" in notes.
  3. Adjusted 3-4 → STOP. Confirm with user before proceeding.
  4. Adjusted <3 → DO NOT APPLY. Inform user and skip.

If the Job Posting has no adjusted score, calculate it first using the scoring table in the job-search skill (which reads config/fit-score.yaml).

Step 0b: Resume-tailoring gate (MANDATORY — NEVER SKIP)

Before ANY application action (form filling, Easy Apply, portal submission, resume upload):

For full protocol (adjusted 7+):

  1. Determine expected tailored resume path: applications/{user-slug}/{company}-{job-slug}/resume.pdf
  2. Check if that file exists for THIS EXACT job posting.
  3. If YES → proceed to Step 1.
  4. If NO → STOP. Run /resume-tailor for this posting first. Create job YAML, run scoring, rewrite bullets, generate .docx and .pdf. Only then proceed.

For longshot protocol (adjusted 5-6):

  1. Use the closest matching generic template (Innovation / Delivery / Generic) — no full tailoring.
  2. Still verify the resume path is a valid generic template, not another company's tailored resume.

HARD RULES:

  • NEVER upload a resume tailored for Company A to an application for Company B.
  • NEVER substitute a "closest match" tailored resume. Each tailored resume contains JD-specific keyword mirroring.
  • Before uploading ANY resume, verify the file path contains the target company name OR is a generic template.
  • For batch applications: apply to adjusted-7+ first (full protocol), then 5-6 (longshot). Don't interleave.
  • Tailor ALL 7+ resumes first, THEN apply all. Don't interleave "apply with whatever is available".

For non-LinkedIn applications (Indeed, Glassdoor, company portals, regional boards)

Step 1: Prepare materials

  1. Read Job Posting record from the database.
  2. Verify tailored resume exists (Step 0 gate).
  3. Draft cover letter if required.
  4. Load standard-answers.yaml for form fields.

Step 2: Navigate

browser_navigate → application URL

Step 3: Understand form

browser_snapshot → capture form structure

Step 4: Fill form

browser_fill_form → populate fields with prepared data

For each field:

  • Name / email / phone → standard-answers.yaml personal section
  • Experience / years → standard-answers.yaml experience section
  • Custom questions → match to common_questions patterns

Step 5: Upload resume

browser_file_upload → upload tailored resume (.pdf or .docx as required)

Step 6: Screenshot before submit

browser_take_screenshot → applications/tracking/{company}-{job}-pre-submit.png

Step 7: Submit

browser_click → submit
browser_take_screenshot → applications/tracking/{company}-{job}-confirmation.png

Step 8: Create database records

  • Create Application record (status: Submitted)
  • Create Interaction record (type: Application Submitted)
  • Update Job Posting status to Applied

For LinkedIn Easy Apply (automated — user-authorized)

Step 1: Prepare materials

  1. Read Job Posting record — get the LinkedIn job ID.
  2. Generate / confirm tailored resume PDF.
  3. Note the resume file path for upload.

Step 2: Navigate to job posting

browser_navigate → https://www.linkedin.com/jobs/view/{JOBID}/

WARNING: do NOT use browser_snapshot on LinkedIn job-detail pages — 500K+ chars, causes loops.

If browser_navigate times out on the snapshot, that's OK — the page is loaded. Proceed with browser_run_code.

Step 3: Open Easy Apply modal and complete form

Use a SINGLE browser_run_code call to handle the entire flow:

async (page) => {
  // Click Easy Apply — MUST use JS click (shadow DOM blocks pointer events).
  await page.evaluate(() => {
    const btn = document.querySelector('button.jobs-apply-button') ||
                document.querySelector('[aria-label*="Easy Apply"]') ||
                document.querySelector('a[aria-label*="Easy Apply"]');
    if (btn) btn.click();
  });

  await page.waitForTimeout(4000);

  // All Easy Apply modal content lives inside #interop-outlet shadow DOM.
  const getDialogState = async () => {
    return await page.evaluate(() => {
      const outlet = document.querySelector('#interop-outlet');
      if (!outlet?.shadowRoot) return { error: 'No shadow root' };
      const dialog = outlet.shadowRoot.querySelector('[role="dialog"]');
      if (!dialog) return { error: 'No dialog' };
      const headings = [...dialog.querySelectorAll('h1,h2,h3')].map(h => h.textContent.trim());
      const buttons = [...dialog.querySelectorAll('button')].map(b => b.textContent.trim());
      const labels = [...dialog.querySelectorAll('label')].map(l => l.textContent.trim());
      const selects = [...dialog.querySelectorAll('select')].map(s => ({
        label: s.closest('.fb-dash-form-element')?.querySelector('label')?.textContent?.trim(),
        options: [...s.options].map(o => o.text)
      }));
      return { headings, buttons, labels, selects };
    });
  };

  const clickButton = async (btnText) => {
    await page.evaluate((text) => {
      const outlet = document.querySelector('#interop-outlet');
      const dialog = outlet.shadowRoot.querySelector('[role="dialog"]');
      for (const btn of dialog.querySelectorAll('button')) {
        if (btn.textContent.trim() === text) { btn.click(); break; }
      }
    }, btnText);
    await page.waitForTimeout(2000);
  };

  let state = await getDialogState();
  const steps = [];

  for (let step = 0; step < 8; step++) {
    state = await getDialogState();
    steps.push(state.headings);

    if (state.buttons.includes('Submit application')) {
      return JSON.stringify({ status: 'ready_to_submit', steps, state });
    }

    if (state.headings.some(h => h.includes('Resume'))) {
      const fileInput = await page.evaluateHandle(() => {
        const outlet = document.querySelector('#interop-outlet');
        const dialog = outlet.shadowRoot.querySelector('[role="dialog"]');
        return dialog.querySelector('input[type="file"]');
      });
      if (fileInput.asElement()) {
        await fileInput.asElement().setInputFiles('RESUME_PATH_HERE');
        await page.waitForTimeout(2000);
      }
    }

    if (state.buttons.includes('Next')) {
      await clickButton('Next');
    } else if (state.buttons.includes('Review')) {
      await clickButton('Review');
    } else {
      break;
    }
  }

  return JSON.stringify({ status: 'navigated', steps, finalState: state });
}

CRITICAL:

  • Replace RESUME_PATH_HERE with the actual absolute path to the tailored PDF.
  • The function navigates through all steps and stops at "Submit application".
  • Take a screenshot BEFORE calling submit.
  • Then call submit in a separate browser_run_code.

Step 4: Screenshot before submit

browser_take_screenshot → applications/tracking/{company}-{job}-pre-submit.png

Step 5: Submit

async (page) => {
  await page.evaluate(() => {
    const outlet = document.querySelector('#interop-outlet');
    const dialog = outlet.shadowRoot.querySelector('[role="dialog"]');
    for (const btn of dialog.querySelectorAll('button')) {
      if (btn.textContent.trim() === 'Submit application') { btn.click(); break; }
    }
  });
  await page.waitForTimeout(3000);

  const result = await page.evaluate(() => {
    const outlet = document.querySelector('#interop-outlet');
    const dialog = outlet?.shadowRoot?.querySelector('[role="dialog"]');
    const text = dialog?.textContent || document.body.textContent.substring(0, 500);
    const success = text.includes('Application submitted') || text.includes('applied');
    return { success, text: text.substring(0, 200) };
  });

  await page.goto('about:blank');
  return JSON.stringify(result);
}

Step 6: Post-submit screenshot + database records

browser_take_screenshot → confirmation
  • Create Application record (status: Submitted, method: Easy Apply).
  • Create Interaction record (type: Application Submitted).
  • Update Job Posting status to Applied.

Standard Answers

The agent reads applications/resumes/data/{user-slug}/standard-answers.yaml for form fields. The schema includes:

  • personal: name, email, phone, location, linkedin
  • experience: years_total, current_title, current_company
  • work_authorisation: status, sponsorship_required (boolean)
  • availability: start_date, willing_to_relocate (regions)
  • certifications[]
  • languages[]
  • common_questions: question → answer pairs for typical screening questions

DO NOT hardcode standard answers in this skill — they're per-user.

Cross-Workstream Integration

Before applying:

  1. Does the company exist in target_companies? → link.
  2. Does the company have networking contacts? → flag as "Warm Application".
  3. Is there an active Sales Pipeline entry? → mention in cover letter.
  4. Should the user reach out to the contact first? → recommend if temperature > Cold.

Anti-Loop Rules (CRITICAL)

  1. NEVER use browser_snapshot on LinkedIn job-detail pages — context overflow + retry loops.
  2. NEVER use browser_navigate to LinkedIn job pages expecting a usable snapshot — it will timeout or return massive data.
  3. ALWAYS use browser_run_code with page.evaluate() for ALL LinkedIn interactions — bypasses shadow DOM, returns only what you extract.
  4. If browser_navigate times out on LinkedIn, the page is probably loaded — proceed with browser_run_code.
  5. Max 2 retries on any single Playwright action — if it fails twice, the approach is wrong. Stop and try a different strategy.
  6. If Easy Apply modal doesn't appear after 2 attempts, the posting may not have Easy Apply. Flag as manual and move on.
  7. After completing an Easy Apply submission, navigate to about:blank — prevents stale LinkedIn DOM bloating subsequent operations.

ATS-Specific Patterns

These are observed patterns. Add new ones as you encounter them; keep them generic.

SmartRecruiters

  • Rejects em-dashes and semicolons in text fields — sanitise before pasting.
  • Phone field: must explicitly click the country-code dropdown option.

Greenhouse.io

  • Use browser_run_code to batch combobox selections efficiently.

Phenom People

  • Simple single-page forms, straightforward to automate.

Oracle HCM

  • Comboboxes use [role="gridcell"]; citizenship uses nationality names (e.g., "Canadian").

Company portals (general)

  • Always browser_snapshot before filling.
  • Always browser_take_screenshot before and after submit.
  • Save all screenshots to applications/tracking/{company}-{job-slug}-{step}.png.

Rules

  1. LinkedIn Easy Apply is AUTHORIZED for auto-submit — use the shadow-DOM pattern above.
  2. ALWAYS screenshot before submission on all platforms.
  3. ALWAYS check for warm-application opportunity before applying.
  4. ALWAYS create both Application + Interaction records after submission.
  5. NEVER retry the same failing Playwright action more than twice — change approach.
  6. Save all screenshots to applications/tracking/.
  7. Parallel resume prep via sub-agents for batch applications (3-5x faster).

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.