nail-performance-review

Remember your achievements when the end-of-year review comes around.

Install it or follow the AI setup guide. If you're installing from a local Claude Code chat, start with the slash command below.

Who is it for?

For whom	What it helps with
Individual contributors	Keep track of the work you do with AI tools, whether you’re an engineer, designer, product manager, accountant, or another role. It fits best when much of your work happens through AI agents or chats on your laptop.
Managers and team leads	Recommend it to your team so they document achievements and impact as they go. Make review season easier when people come to you with their contributions already collected.

Your records stay in a folder on your laptop. Your chosen AI provider may process the notes your AI chat reads.

Why I created it

I think we all want to get what we deserve based on our work and make our lives easier by collecting the evidence.

I created this plugin after spending about a week talking to my AI agent, trying to find the docs, projects, tickets, and chats I had worked on.

Just using AI at the end of the year isn’t enough. Chats can disappear, documents get lost, and some work never gets a ticket. Remember helping Kate learn a new tool during a quick 1:1 eight months ago? Did you ticket it? If not, that’s an example of mentoring and collaboration you might forget to mention in your review.

How does it work?

Step	What happens
1. Start	Install the plugin, then say “Help me set up my performance record.” You can also choose `nail-performance-review-setup` from your AI tool’s skill menu.
2. Give it some context	Start with your role, review period, and goals. Upload files or paste the text. Add development goals, promotion interests, your company’s career ladder, past reviews, or manager feedback when you have them. You don’t need all of these to begin. Use only information you’re authorized to store and share with your AI provider.
3. Keep working	In an enabled workspace, your AI can save quick fixes, useful findings, and work in progress to a private inbox. Later, choose what to include in your review. You can also say “Record what we achieved.”
4. Check your priorities	Your AI chat checks the project against your saved yearly goals. If nothing matches, it asks whether you want to connect it to a goal, add an agreed goal, or work on it anyway. If it doesn’t benefit your priorities, is it something you want to spend your time on?
5. Collect the impact	When the work in your chat suggests the project is nearly finished, your AI chat considers whether there are useful KPIs, before-and-after measurements, or feedback to collect. If the shared reminder limit allows, it asks what to collect and saves results with supporting links.
6. Help with your growth	As a project wraps up, your AI chat checks your development goals and company expectations. It might suggest a broader demo, documentation, or applying the learning to another project.

Goal-alignment, impact, and growth reminders share a limit of one per session and two per seven days across your configured tools. You can also snooze or dismiss them.

A career ladder describes what your company expects at different levels. You don’t need an upcoming promotion to start developing those skills. From what I’ve seen, people often show the skills of the next role before they get the title.

That’s it. Keep doing your work, and keep the useful evidence along the way.

The plugin doesn’t require an extra AI subscription. Use it with your current local Codex, Claude Code, Cursor, or Gemini CLI setup. Web chats use a manual handoff. Let me know what’s missing by opening an issue, sending a PR, or emailing stoiancee@gmail.com.

Example: migrate Google Cloud to AWS while growing from Senior toward Staff. In a workspace you enable for tracking, your AI chat uses your saved goals and the work you discuss to spot milestones and useful follow-ups. Measurements come from you or tools you authorize.

Install it

Using local Claude Code? Install Git and Node.js 22 or later. In Claude Code 2.1.275 or later, enter:

/plugin install nail-performance-review --marketplace yaro-work/nail-performance-review

Choose User scope, open a new chat, and say “Help me set up my performance record.” If you used the terminal installer before, follow the switching steps.

Your AI tool	Install from GitHub
Gemini CLI	`gemini extensions install https://github.com/yaro-work/nail-performance-review`
Codex CLI	`codex plugin marketplace add yaro-work/nail-performance-review`, then `codex plugin add nail-performance-review@yaro-work-nail`
Cursor	In Customize → Plugins, import `https://github.com/yaro-work/nail-performance-review` from GitHub, then select Install.

Start a new chat after installing. See the native install guide for hook trust and switching from the terminal installer.

Cursor can import Claude Code hooks on some setups. If you use both tools, check for imported hooks before adding another copy.

Prefer help from your AI? Give this README or repository link to your local AI coding tool and paste:

Help me install nail-performance-review. Follow the AI setup guide. Give me one next step at a time.

Prefer a few commands? Install Git and Node.js LTS, version 22 or later if you don’t have them. Node.js is the program that runs this plugin. Reopen your terminal after installation.

Download the plugin and create your record. By default, notes go in the performance folder inside your home folder.
```
git clone https://github.com/yaro-work/nail-performance-review.git
cd nail-performance-review
node bin/nail.mjs init
```
Automatic tracking starts on for all folders. If you prefer to save work only when you ask, run node bin/nail.mjs init --manual instead.

Run the install command for each AI tool you use. These are alternatives; you don’t need all four.

Your AI tool	Run this command
Codex	`node bin/nail.mjs install --target codex`
Claude Code	`node bin/nail.mjs install --target claude`
Cursor	`node bin/nail.mjs install --target cursor`
Gemini CLI	`node bin/nail.mjs install --target gemini`

Run node bin/nail.mjs doctor. Complete your AI tool’s trust or enablement steps, then start a new chat and say “Help me set up my performance record.”

Enable automatic tracking

New setups track all folders by default, including subfolders and new projects. One private record is shared across your local AI tools. You can switch to selected folders or exclude a folder and all its subfolders.

For manual capture, say “Switch to manual tracking” or run node bin/nail.mjs tracking --manual. This turns off automatic capture in every folder. You can still ask your AI to save work.

Existing setups keep their tracking choices. To switch an existing record to all folders, say “Enable automatic tracking in all folders” or run the command below. Replace work if your record has a different name.

node bin/nail.mjs tracking --enable --all --name work

Complete any hook trust steps and open a new chat. To track just one project or exclude a folder, use the tracking settings.

Try it with real work. A quick fix, a useful diagnosis, or work you’ve started can go into your private inbox, even without a measured outcome. Ask “Show my saved candidates.” Choose what to include in your review or leave out. If your AI missed something, say “Save this to my performance record.” Tracking checks

After setup or an update, ask “Is my tracking working?” to check this folder and see when each AI tool last saved a candidate.

Use your goals and one-on-one notes

Share a goals document or 1:1 progress notes and say “Use this whole document to fill in my goals and achievements for this review period.” Last year’s self-review and your manager’s review or guidance can help identify what to focus on now. Share only material you’re allowed to use with your AI tool.

Prefer Google Docs or other links your AI can open. Manager 1:1s default to weekly checks. Other tracked links default to every 30 days. Choose a different interval or a one-time review. Due checks and retries run when you use an eligible chat in the selected workspace. Ask “What changed in my linked docs?” to see what was checked, saved, or will retry. Set up document checks

Update it

After two weeks, the startup hook checks the public version on GitHub and tells you when an update is available. It checks only in workspaces with tracking enabled. You choose when to install it. Check now or turn off notices

If you installed through Claude Code's plugin browser, open /plugin, select Installed, then choose Update now for nail-performance-review.

For other native installs, use your AI tool’s updater.

If you installed from a clone, close your AI tools and run one command in the plugin folder. It updates every adapter you installed from the terminal, including Codex, Claude Code, Cursor, and Gemini CLI.

node bin/nail.mjs update --pull

Reopen your AI tools and start a new chat. Your records, tracking choices, and installation modes stay in place. If an older copy rejects --pull, use the one-time upgrade steps.

Prepare a review

Ask “Review my week and prep my 1:1.” Pick any new notes you want to use. Your AI drafts up to four short bullets grouped by goal or OKR.

Ask for an annual, mid-year, or custom review. The plugin gathers the saved evidence; your AI chat drafts a concise self-review with your contributions, collaborators, goals, and links. Ask for an offline HTML overview if you’d rather browse your achievements that way.

Before you rely on it

This is a personal recordkeeping tool, not an employee-monitoring or employment-decision system. Check generated claims, credit, and evidence before sharing. Missing records do not establish poor performance.
Capture depends on your AI chat following the tracking instructions and having permission to save files. It doesn’t guarantee every achievement is saved. Check important milestones yourself.
Your AI chat interprets progress from the current conversation. The plugin doesn’t independently monitor projects or dashboards, read unrelated chat histories, or invent missing measurements.
Use your company’s expectations first. The career examples are editable starting points, not promotion rules.
Real Codex chats have saved achievements, corrected credit, prepared reviews, remembered a priority choice, and asked about project outcomes. Claude Code, Cursor, and Gemini CLI have automated installation and context tests; we haven’t verified these conversations in those tools.
Records are local Markdown files, without encryption from the plugin. Your AI provider’s policies still apply to the notes it reads. Hosted chats without local file access need a manual handoff.

Provided as is under the MIT license, without warranties. No promotion or compensation outcome is guaranteed. Read the short use notice and privacy guidance.

See these docs only if you want to understand how this plugin was tested and developed in depth. Feel free to give the repository to your AI and ask questions that matter to you: what we tested, the repeatable evaluation, and its results.