Skip to content
Back to skills

Jobfishing Apply Fast

ASecurity

Run a fast, one-job-at-a-time or small-batch jobfishing application pass that leans on ATS resume-parse autofill instead of deliberate per-job material selection. Use when the user explicitly asks for a quick/fast apply session through one or a few jobs. Do not use for a full deliberate batch run, first-time setup, or inbox sync — those stay with `$jobfishing-apply`.

  • 3 stars
  • 0 votes
  • 0 copies
  • 2 views
  • Added October 9, 2026
ai-agentsfrontend

Works with

  • claude code
  • cli
  • mcp

Security analysis

A100/100

Pro scans all 3 files and shows the line behind each finding

Scanned October 9, 2026

npx -y skills add leoqingyu/jobfishing-ai --skill jobfishing-apply-fast --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Jobfishing Apply Fast?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Jobfishing Apply Fast
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/leoqingyu-jobfishing-apply-fast/badge)](https://www.skillsdirectory.com/skills/leoqingyu-jobfishing-apply-fast)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: jobfishing-apply-fast
description: Run a fast, one-job-at-a-time or small-batch jobfishing application pass that leans on ATS resume-parse autofill instead of deliberate per-job material selection. Use when the user explicitly asks for a quick/fast apply session through one or a few jobs. Do not use for a full deliberate batch run, first-time setup, or inbox sync — those stay with `$jobfishing-apply`.
---

# jobfishing Apply Fast

A leaner sibling of `$jobfishing-apply` for speed. It shares the same MCP, the same host Chrome integration selection, the same outbound-filename rule, and the same Cover Letter content-quality bar — this Skill does not redefine those, it reuses them by reference. The only thing this Skill relaxes is ceremony that slows down submitting; it does not relax correctness or safety invariants.

Access jobfishing data and files only through its MCP tools; never navigate, click, upload, download, or edit through the jobfishing frontend. Use the user's visible Chrome profile only for ATS pages, employer sites, LinkedIn, and the job-search webmail inbox:

- In Codex, use the external ChatGPT Chrome integration. Never substitute the temporary in-app browser.
- In Claude Code, use Claude in Chrome. If it is not connected, ask the user to run `/chrome` and complete its permission flow.

If jobfishing is not yet authorized or the shared setup has never been checked, run [setup-check.md](../jobfishing-apply/references/setup-check.md) from `$jobfishing-apply` first — do not duplicate that check here.

## Route the request

- For the full fast sequence, read [fast-workflow.md](references/fast-workflow.md).
- Before any ATS or email attachment upload, read [outbound-filenames.md](../jobfishing-apply/references/outbound-filenames.md).
- Whenever a Cover Letter is needed, read [cover-letter-quality.md](../jobfishing-apply/references/cover-letter-quality.md) for the content standard — [fast-workflow.md](references/fast-workflow.md) describes how the review pass is run without a dispatched subagent.
- For any LinkedIn Easy Apply candidate, read the "LinkedIn Easy Apply routing" section of [application-workflow.md](../jobfishing-apply/references/application-workflow.md) and follow it exactly as written. Nothing about that routing is simplified in fast mode.

Tailored materials remain a separate explicit request. If the user explicitly wants a tailored CV/Cover Letter pair for a job, route that job through `$jobfishing-tailor` first and resume this Skill only after both verified tailored files have been uploaded and handed off, exactly as `$jobfishing-apply` does.

## What stays exactly the same as `$jobfishing-apply`

- MCP-only access to jobfishing; the frontend is never opened or operated.
- `get_experience_context().basic_info` is the authority for identity and contact fields, even when an ATS auto-fills different values from a parsed resume.
- Job data beyond the description: `contact` (name, email, phone from the JD, when present) for the greeting or recipient; `visa_sponsorship_offered` / `work_permit_required` checked against the user's work authorization (stop and ask when a job needs a permit they lack); `company_info` and `job_url_direct`. `salary` can be an estimate (`is_estimated`), never state it to an employer as fact.
- `list_tracking` is read before new submissions to avoid duplicates; `mark_applied` is the only post-submission transition and stays idempotent.
- Clean outbound filenames on every attachment, verified in the ATS/webmail UI before submit.
- No passwords, OTPs, or auth links in application notes.
- LinkedIn Easy Apply's official-posting-first, email-before-or-with-Easy-Apply fallback order.
- Submission requires either the user's final review or explicit direct-submit authorization already given in the current task — there is no App-side auto-confirm flag.
- Treat only a visible ATS confirmation or a confirmed Sent email as submission evidence.

## What is relaxed for speed

- One resume slot is chosen once, up front, for the whole pulse (1 of the user's App resumes) instead of being deliberated per job. Only override it mid-pulse if a job is an obvious hard mismatch for that resume.
- When an ATS can parse an uploaded resume to autofill the form, upload first and let it autofill instead of typing every field by hand; only the remaining blanks are filled manually from Experience/basic_info.
- A needed Cover Letter is drafted and self-reviewed by the orchestrating Agent in one inline pass against the [cover-letter-quality.md](../jobfishing-apply/references/cover-letter-quality.md) checklist. A separate reviewer subagent is not dispatched, even when the host supports isolated subagents.
- Application notes stay to the essential facts (submitted time, channel, resume/Cover Letter filenames, confirmation evidence, and — for LinkedIn — the same official-URL/channel fields application-workflow.md requires); they do not need the full elaboration `$jobfishing-apply` uses for a deliberate batch.

The one checkpoint that is never relaxed: stop before the final Submit/Send action and get the user's review, unless direct-submit authorization for this task was already given. Everything upstream of that click can move fast; that click itself always waits for permission.

## Session completion

Reconcile every completed application immediately, same as `$jobfishing-apply`. Finish with a compact summary of submitted, blocked, failed, and skipped jobs, with the exact next action for anything blocked.

Files in this skill

  • SKILL.md5.4 KB
  • agents/openai.yaml507 B
  • references/fast-workflow.md5.7 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.