career-skills

Career Skills

A collection of Agent Skills to help you manage and grow your career. Works with any agent that supports the Agent Skills standard, including Claude Code, GitHub Copilot, Cursor, Gemini CLI, and others.

Skills

Skill	What it does
career-review-cycle	Set goals, log work, generate brag docs, prepare for reviews and 1-on-1s, build a promotion case

Installation

Any Agent Skills compatible agent

npx skills add p-m-p/career-skills

Claude Code

Add this repository as a plugin marketplace, then install:

/plugin marketplace add p-m-p/career-skills
/plugin install career-skills@p-m-p

career-review-cycle

Covers the full year cycle: set goals at the start of the year, log work as you go, then synthesise that log into career documents when you need them. Just talk to your agent naturally — it activates the skill and routes to the right capability based on what you ask for. The more consistently you log, the less effort everything else takes.

1. Set your goals

At the start of a review cycle, tell your agent what you're working toward. The agent will help you shape goals that cover both your core role and areas where you want to grow or increase your visibility — using OKR thinking as a guide: a clear objective and measurable outcomes that tell you when you've got there.

"Help me set my performance goals for this year. My team is focused on reliability and I want to grow into more of a technical lead role."

Goals are saved to your vault and referenced throughout the rest of the cycle.

2. Log as you go

The most important habit is capturing work while it's fresh. After a meaningful piece of work — shipping something, unblocking a team, running a cross-functional meeting — take thirty seconds to log it.

"Log that I finished the payment retry logic today. It was blocking the iOS team and we'd had three incidents because of it."

The agent links the entry to your goals and prompts for the outcome if you haven't mentioned one. Over time this builds a precise record that's almost impossible to reconstruct from memory at review time. Log entries follow STAR principles — capturing the situation, what you did, and the result — so they're already shaped for use in review documents later.

3. Prep for 1-on-1s

Before each manager 1-on-1, ask your agent to pull together what you've been working on. It reads your recent logs and surfaces wins, blockers, and anything that needs a conversation — including goals that are drifting off track.

"I've got a 1-on-1 tomorrow, what should I raise?"

This keeps your manager informed without you having to remember everything, and creates a habit of surfacing asks — decisions you need, introductions, feedback — rather than just reporting status.

4. Generate a brag doc each quarter

Every quarter or half-year, synthesise your logs into a brag document. This is a structured record of your impact, grouped by goal, with raw activity reframed into outcomes. It's the source material for everything else.

"Write up my accomplishments for Q2."

The agent flags gaps where coverage is thin and asks for any outcomes you didn't capture at the time. The result is a document you could share with your manager or use verbatim in a self-assessment.

5. Prepare for performance reviews

When review season arrives, your brag doc does most of the work. The agent uses it — alongside your goals — to write a self-assessment in the language your company uses, with accomplishments framed using STAR (Situation, Action, Result) so every claim is backed by evidence.

"My performance review is next week, help me write my self-assessment."

If your company has a specific review form or rubric, share it and the agent will adapt the output to match.

6. Build a promotion case

When you're ready to make a case for promotion, the agent maps your strongest accomplishments to the criteria for the next level. It shows where the evidence is strong, where it's thin, and what work would close the gaps before the next cycle.

"I want to make a case for promotion to staff engineer. Here's our leveling guide."

---

All output is plain Markdown stored in a vault folder of your choosing — works great with Obsidian.