TL;DR: I set up a Claude Code skill that checks in with me about my workday, asks follow-up questions, and saves a structured markdown file I can use as promotion evidence. Here's why it works, and how to build one in about five minutes.
May 6th
On May 6th I had an energy level of 2 out of 5.
I got my Claude Certified Architect exam score back that day: 717 out of 1000. I needed 720. I missed it by three points.
Four lines down in the same entry, my manager had told me: "your leadership is being felt around Artium. You're making a good impact."
Here's the thing about that day: the bad number is vivid and self-evident. 717. Three points short. That number was going to live in my head rent-free for weeks. But the recognition? That quietly evaporates. Left to memory, May 6th is the day I failed the exam by three points. On the page, it's also the day my manager told me my leadership was landing across the company.
The entry keeps the thing I'd lose otherwise.
The Problem With Memory
I've been bad at this for years. At performance review time, I'd stare at a blank document trying to remember what I'd actually done. I'd come up with four things instead of forty. My manager would advocate for me based on what she happened to see, which was never the full picture.
The thing is, I did good work. I just didn't capture it.
A few years ago I tried to solve this with Google Forms, a structured form I'd fill out at the end of each day that fed into a spreadsheet. It worked, kind of. The data was there, but it felt like homework. The form didn't ask follow-up questions. It didn't notice when I was being vague. I had to go somewhere specific to fill it out. And when review time came, I had to go back somewhere else to compile everything, figure out what mattered, and assemble it into something coherent. The friction wasn't just the daily entry. It was the whole chain: capture, retrieve, synthesize, present. I was on my own at every step.
So I built something better.
What I Built
I have a git repository called daily_self_evals. Inside it, every workday gets its own markdown file:
entries/
└── 2026/
└── 03/
└── 2026-03-02.md
Each file has YAML frontmatter (date, day, energy level, tags) and sections for what I shipped, the impact of that work, wins, challenges I overcame, recognition I received, and what's next. Here's a representative example of what an entry looks like:
---
date: 2026-02-12
day: Thursday
week: 7
energy: 4
tags: [shipped, cross-team, recognition]
---
# Daily Eval — Thursday, February 12, 2026
## Shipped
Wrapped up the LLM observability integration. Multi-turn conversation
history is now captured and agent tool calls are visible for evaluation.
This had been a gap in the platform for a while.
## Impact
The team called it out as a major win in our retro. Other engineers
can now actually see what the agent is doing across a full conversation,
not just at individual turns.
## Wins
Got the company spotlight award today. Wasn't expecting it.
## Recognition
My manager flagged it in the all-hands. A few people I don't work
with day-to-day sent messages saying they'd been waiting for this.
The Skill
In Claude Code, you can build a custom skill: a saved workflow with a description that tells Claude when to use it. I built one for journaling, and the nice part is I don't invoke it with any special syntax. I just talk to Claude Code the way I'd talk to a coworker:
"Hey, let's make an entry for today."
"Can you record this win I just had?"
Claude recognizes that these match the journaling skill and runs it. (You can also invoke it manually with /daily-eval if you prefer typing a command, but I never do.)
When it runs, Claude first checks if an entry already exists for today. If it does, it reads what's there and asks what I want to add. If it doesn't, it starts fresh. Either way it appends and never overwrites, so I can log something at 2pm and add more at the end of the day, and it all lands in the same file.
Then it asks me questions, one at a time, conversationally:
"What did you ship or move forward today, even if it was small?"
"What effect did that have on your team or the product?"
"Did anyone thank you or give you good feedback?"
It's not a form. It feels more like a five-minute debrief with a colleague who's genuinely curious. When we're done, it writes or updates the file and tells me where it saved it.
Why the Conversational Format Matters
The STAR framework (Situation, Task, Action, Result) is embedded in the question flow without being named. When Claude asks "what effect did that have?" after I describe a task, I think harder. I give better answers than I would in a form. By the time I'm done, each entry reads like something I could hand to a hiring manager or drop into a performance review, because the prompts shaped it that way from the start.
What You Can Do With the Files
Everything is plain markdown with YAML frontmatter. No app. No database. No lock-in.
That means you can:
- Export to Word or PDF:
pandoc entries/**/*.md -o accomplishments.docx - Import the whole folder into Notion or Obsidian
- Ask Claude to summarize three months of entries before a performance review
- Paste any entry directly into a promotion packet or 1:1 doc
- Push it to a private GitHub repo for a permanent, timestamped record
Because the YAML frontmatter includes structured fields like energy level and tags, you can also ask Claude to generate charts and analysis across your entries. Energy level over time. Which weeks had the most wins. How often you shipped versus how often you were blocked. The Google Forms version of this system fed into a spreadsheet that could do exactly this kind of visualization. The Claude Code version can do the same thing, just by asking.
One More Use: When Someone Asks You for Feedback on a Colleague
Peer feedback requests always arrive at the worst time. Someone asks you to fill out a survey about a teammate you worked closely with three months ago, and you stare at the form trying to remember anything specific.
If you've been logging your interactions, you don't have to remember. You can ask your journal: "When did I work closely with [name] and what did I notice?" You'll have the actual moments, the specific things that helped you, the friction you observed. Your feedback becomes something a manager can actually act on, instead of the generic "great to work with, highly recommend" that tells everyone nothing.
The record makes you a better peer, not just a better advocate for yourself.
How to Set This Up
You need Claude Code installed. Then just ask Claude to build it for you. You don't need to create any files manually. Open Claude Code and paste something like this:
"I want to set up a daily work journal to build promotion evidence. Create a
~/daily_self_evals/project with aCLAUDE.md, anentries/directory organized by year and month, and a global skill at~/.claude/skills/daily-eval/SKILL.md. Give the skill a description so you'll recognize when I ask you to make a journal entry or record a win in plain language, and make it invocable as/daily-evaltoo. When it runs, it should check if an entry already exists for the target date before asking questions, ask one question at a time about what I shipped, the impact, wins, challenges, recognition, and what's next, and save entries as markdown files with YAML frontmatter. All entries should be appendable and never overwrite existing content."
Claude will handle every file. You just review what it built, open a fresh conversation, and tell it you want to make today's entry. (Skills load at the start of a conversation, so open a new one after setup.)
What Happens After a Few Months
I now have a body of evidence I've never had before. Specific accomplishments. Measurable impact. Quotes from colleagues. An energy rating that shows me my own patterns over time: what kinds of days light me up, what kinds drain me.
When it's time to make the case for a promotion, or find a role that's more interesting and pays better, I won't be starting from memory. I'll be starting from a record.
And on the days when I miss a certification by three points and want to believe the whole day was a loss, the record knows better.
One last thing: this system isn't finished. I keep tweaking the questions, adjusting what the frontmatter captures, experimenting with how I ask for summaries. That's part of the point. It's a tool I own and can improve, and the more I use it, the better I get at using it. The version I'm running today is better than the one I started with, and the version I'll be running in six months will be even better.
Inspired by Get your work recognized: write a brag document by Julia Evans, and my earlier experiment with Google Forms for self-evaluation.
Top comments (0)