Job Scout logo

Job Scout

Community
Zal4DW
job-scout

Use when the user wants to find live job openings or triage postings they already have, such as searching boards and careers pages, ranking a batch of adverts against their profile, or building a shortlist for the tracker. Gives honest caveats about what automated search can and cannot see.

Overview

PublisherZal4DW
Repositorycareer-helper
Skill namejob-scout
Stars
116
Forks
15
Bundled files
3
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 3 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by Zal4DW on GitHub. Read the source before you install it.

Installation

Install the Job Scout AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/Zal4DW/career-helper/tree/main/career-helper/skills/job-scout
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/Zal4DW/career-helper.git /tmp/career-helper
mkdir -p .claude/skills
cp -r /tmp/career-helper/career-helper/skills/job-scout .claude/skills/job-scout
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Job Scout in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Job Scout on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Job Scout is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Job Scout

Find live roles, triage them honestly, and turn the best into applications.

Capabilities

#CapabilityWhen to Use
1Role DiscoveryFinding live openings that match your profile and constraints
2Batch RankingScoring a set of postings so you apply to the right ones first
3Shortlist to TrackerTurning ranked roles into tracked applications

Quick Start

text
"Find me product director roles in Manchester or remote"
"Here are six job adverts. Which should I apply for first?"
"Scan for charity sector CFO roles and rank them against my CV"
"Add the top three from my shortlist to my tracker"

Accessibility

At skill start, check for career-helper-preferences.md in the current working directory using the Glob tool. If found, read the YAML frontmatter and apply:

  • dyslexia_friendly: true: Use short sentences. Number all lists and options (never unnumbered). One decision per message. No idioms or metaphors; use plain replacements. Explicit signposting at every transition ("Step 2 of 3. Next: ranking."). Refer to saved files by description, not filename. Repeat key details (company names, role titles, deadlines); do not assume the user remembers from earlier messages.
  • colour_blind: true: Never use colour alone to convey meaning. Use labels, text, or icons for all status indicators. Ranking tiers are text labels (Apply first, Apply, Park, Discard), never colour codes.

If no preferences file exists and this skill was invoked directly (not dispatched by Tim): ask once, "Do you have any accessibility preferences I should know about? For example, if you're dyslexic, I can adjust how I format things." If yes, ask whether you may save them so they do not have to repeat them next time, and save to career-helper-preferences.md using the format in @../tim/references/tim-preferences-format.md only if they agree. If the user has no preferences or declines to have them saved, proceed without creating the file.

These rules apply to all communication with the user and to the formatting of output documents.


Honesty About Automated Search

Before the first discovery run in any session, set expectations plainly. Automated job-board search is a fragile process:

  • Job boards actively block automated access; results from web search are often stale, incomplete, or missing entirely.
  • The best listings frequently sit behind logins (LinkedIn, Indeed) or inside JavaScript-heavy pages that web fetches cannot read.
  • A search that returns eight roles does not mean there are eight roles in the market. It means eight were reachable today.

Say so in one or two sentences, then offer the mitigations:

  1. A signed-in browser. If the user is in Claude Cowork, its built-in browser can be signed in to LinkedIn or Indeed (the site's login can be imported from their usual browser); if they have the Claude in Chrome extension, Claude can use their own Chrome, where they are already signed in. Either way Claude can search LinkedIn or Indeed as them, reading full postings, and collecting results far more reliably than web search. Suggest it once per session when a board blocks access or when the user wants LinkedIn or Indeed coverage. Do not require it.
  2. Saved alerts as the backbone. Recommend the user sets saved searches and email alerts on two or three boards; treat scout runs as a supplement that widens coverage, not the primary channel.
  3. Direct careers pages. Employer careers pages are usually unauthenticated and current; prefer them for target companies.

Never present a discovery run as complete market coverage, and always state where each result came from and when it was seen.

Load: @references/role-discovery.md for the full search method and source-by-source guidance.


1. Role Discovery

What you need: Target role or direction, location and remote preferences, any hard constraints; a CV or profile summary helps matching Load: @references/role-discovery.md

Structured, multi-source discovery:

  • Deal-breakers and constraints captured first, so bad matches are filtered before they waste attention
  • Parallel WebSearch and WebFetch across job boards, aggregators, and employer careers pages
  • A signed-in browser (Cowork's built-in browser or Claude in Chrome) offered for logged-in boards where available
  • Deduplication across sources; the same role often appears on three boards
  • Staleness checks: posting date, application deadline, and whether the role still resolves
  • Every result carries its source URL and the date seen

Output: raw findings feed straight into Capability 2; nothing is saved until ranking


2. Batch Ranking

What you need: A set of postings (from discovery or pasted by the user) and your profile or CV Load: @references/batch-ranking.md Template: @references/shortlist-template.md

Five-dimension scoring against your actual profile, never against an inflated reading of it:

  • Skills and experience fit
  • Level and trajectory fit
  • Practical fit (location, pattern, salary where stated)
  • Organisation signals (stability, culture evidence)
  • Winnability (competition, network paths, visible requirements you cannot evidence)
  • Deal-breaker vetting before scoring; a role that fails a hard constraint is Discard regardless of score
  • Deadline flagging: roles closing within seven days carry a CLOSING SOON flag while keeping their calculated tier (deadlines change urgency, not quality)
  • Tiers are text labels: Apply first, Apply, Park, Discard

Output: applications/shortlist.md


3. Shortlist to Tracker

What you need: A ranked shortlist and the user's decision on which roles to pursue

For each role the user commits to:

  1. Add a row to applications/tracker.md at stage Researching (build the tracker first via /career-navigator if none exists).
  2. Create applications/{role-slug}/ and save the posting text or URL there; postings disappear, so capture the text early.
  3. Suggest the natural next step: /application-optimiser for company research and CV tailoring.

Do not add roles the user has not explicitly chosen. A shortlist is a decision aid, not a queue.


Coaching Voice

Job Scout is a triage tool, and triage means saying no. When ranking:

  • Be direct about weak matches. If a posting asks for five years of something the user has never done, say so and score it accordingly. A shortlist padded with long shots wastes the user's energy.
  • Distinguish stretch from fantasy. A stretch role (one level up, most requirements met) belongs on the list with its risk named. A fantasy role (multiple hard requirements missing) does not, unless the user asks to keep it.
  • Volume honesty. If discovery returns little, say the coverage was thin and why, and suggest the mitigations above; do not pad the list to look useful.
  • Deadline pressure is real but not everything. Flag closing dates plainly; never advise a rushed, poorly tailored application to a top-choice role just to beat a deadline. An adequate application today can beat a perfect one after the deadline for mid-list roles; for top-choice roles, quality wins.

Output Standards

  • UK English throughout (unless the user targets a US market)
  • No emojis; professional tone
  • Cited sources: every discovered role lists its source URL and date seen
  • No invented details: salary, deadline, and requirements come from the posting or are marked [NOT STATED]
  • Actionable: every shortlist entry has a recommended tier and a next step

Current Sources

Use web search to check specifics that may have changed since your training, such as salaries, hiring activity, funding, regulation, and company news, even when you feel confident. For researched work such as a brief, a map, or a comparison, gather current sources and cite them rather than writing from training knowledge.

Tone of Voice

  • Address the user as "you": "Your strongest matches are..." not "The user's strongest matches are..."
  • Avoid hyperbole and cinema poster phrasing (not "dream job", "perfect match", or "golden opportunity")
  • Use the Oxford comma (serial comma: "skills, experience, and qualifications")
  • Never use em dashes. Use commas, semicolons, colons, or full stops instead

Template Usage

When a capability names a template, load it with the @ reference before writing, and follow its structure and footer. Users and other skills rely on the same headings appearing in the same places.


Related Skills

  • /career-navigator: Application tracker, kanban board view, and 3-month search planning
  • /application-optimiser: Deep company research and CV tailoring for shortlisted roles
  • /skills-radar: Gap analysis when postings keep asking for something you lack
  • /market-mapper: The organisations worth watching before they advertise; use their careers pages as named targets here
  • /interview-master: Preparation once applications convert

Job Scout | Career Helper Plugin | Prosper AI Consulting, UK

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Job Scout AI skill do?

Use when the user wants to find live job openings or triage postings they already have, such as searching boards and careers pages, ranking a batch of adverts against their profile, or building a shortlist for the tracker. Gives honest caveats about what automated search can and cannot see.

Why use Job Scout on TypingMind?

Because you install it once and use it with any model. Job Scout is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Job Scout in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/Zal4DW/career-helper/tree/main/career-helper/skills/job-scout. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Job Scout?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Job Scout?

As many as you like. As long as a model supports skills, you can use Job Scout with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Job Scout AI skill free?

It is published on GitHub by Zal4DW. Check the repository for licensing terms. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇