/upskill

Analyze skill gaps across your tracked jobs (aggregate) or one posting (targeted), then build a prioritized, web-sourced learning plan and save a report.

Boost this command

suraj-davariya@suraj-davariya

GitHub

From plugin

suraj-davariya-ai-job-search

236 skills2 agents6 commands

Install

$ npx -y skills add suraj-davariya/ai-job-search --agent claude-code

How it fires

How this command gets triggered: by you, by Claude, or both.

Fires itselfClaude auto-loads it when your prompt matches the work.
You can call itInvoke it directly when you want it.
Slash command/upskill

Context preview

What this command does when you run it.

Analyze skill gaps across your tracked jobs (aggregate) or one posting (targeted), then build a prioritized, web-sourced learning plan and save a report.

Command definition

upskill.md

description: Analyze skill gaps across your tracked jobs (aggregate) or one posting (targeted), then build a prioritized, web-sourced learning plan and save a report.
argument-hint: "[job url | pasted posting | (blank for aggregate)]"
allowed-tools:
  - Read
  - Write
  - Edit
  - Bash
  - Glob
  - Grep
  - WebSearch
  - WebFetch

/upskill — Skill Gap Analysis

> **Spec:** REQ-3001–3011, ARCH-0040 (Career Development), ARCH-0008 > **Backed by:** the `career-development` Plane-1 skill > **Used by:** the dashboard Upskill surface (reads the `upskill/report-*.md` files this produces)

Activate the **`career-development`** skill and run its full workflow:

1. **Pick the mode** (REQ-3001):

`/upskill` (no argument) → **aggregate** mode — analyze every job in

`job_search_tracker.csv`. If the tracker is empty, say so and suggest targeted mode instead of writing an empty report.

`/upskill <url>` (or pasted posting text) → **targeted** mode — analyze one posting.

If the URL can't be fetched, ask for a paste (DEC-011); never abort. 2. **Load data** (REQ-3002, REQ-3003) — candidate profile always; tracker + most recent previous report in aggregate mode; the fetched/pasted posting in targeted mode. 3. **Pass 1 — hard-skill diff** (REQ-3004) — required/preferred skills vs. profile, fit-weighted in aggregate mode; generous matching removes skills already held. 4. **Pass 2 — synthesis** (REQ-3005) — domain / soft / tooling / credential gaps. 5. **Gap heatmap** (REQ-3006, REQ-3011) — Critical/High/Medium/Low table printed first; Low gaps listed but not planned. 6. **Learning plan** (REQ-3007) — 2–3 **real, web-searched** resources per Critical/High gap (and Medium when <5 gaps total), with study direction and time estimate. Never fabricate resources (ARCH-0007). 7. **Study order** (REQ-3008) — dependency-aware numbered sequence with total time. 8. **Save the report** (REQ-3009, REQ-3010) — always write `upskill/report-YYYY-MM-DD.md` (aggregate) or `upskill/report-YYYY-MM-DD-<company>-<role>.md` (targeted), including the since-last delta in aggregate mode.

Full behavior lives in `.claude/skills/career-development/SKILL.md`.

Read it on GitHub ↗

Ships withsuraj-davariya-ai-job-search

An AI job-search assistant that finds postings, writes tailored CVs and cover letters, and compiles them to print-ready PDFs — all on your own machine.

Get the whole plugin

Stats

Python TypeScript JavaScript Node React

+7 tags

Next.js Tailwind Bash Git GitHub LaTeX Anthropic

Stars

Forks

Maintained

Maintenance

MDX

Language

MIT