- Home
- Skills
- Education & Learning
- Learning Log
Learning Log
Keep a short record of the topics you worked through with an AI agent. Substantial learning gets an entry; quick clarifications stay out.
Free
Works with the AI Agents you already use
Learning Log
Example session with this skill installed
Use Learning Log during this CNN study session. Save to learned-log.md.
We traced a 1x28x28 input through sixteen 3x3 filters with stride 1 and padding 1, then a 2x2 pooling layer. I corrected my assumption that pooling reduces channels: the result is 16x14x14. I also calculated stride 2 without padding as 16x13x13, and worked out why a larger image increases compute without increasing the shared filter parameters. I understand the forward-pass shapes; I haven't covered backpropagation.
- Read your context and instructions
- Compiled the learning log
Synthetic study session, evaluated with this skill. This is the complete saved log:
# Learned Concepts
- [x] **CNN forward-pass reasoning** | ml | 2026-11-14 | Traced how padding, stride, filters, and pooling affect tensor shapes while shared weights keep parameter counts independent of image size.
One forward-pass topic covers the connected work on shapes, shared filters and pooling. Backpropagation wasn't included because the session hadn't established that understanding.
In the isolated replay check, presenting the same exchange again left one entry in the file. Routine capture is silent; the file is shown here to make the result visible. Short padding and frozen-field clarifications were tested separately and produced no entries.
Connects securely to your tools. The creator never sees your data.
About this skill
A good learning session rarely stays tidy. You trace an example, get a detail wrong, debug it, then work out why it behaves that way. The conversation moves on.
Learning Log keeps a short record of those sessions. It waits for substantial work on a connected topic and looks for understanding in your own reasoning. A correct answer to a small question doesn't automatically earn an entry.
Working through a CNN forward pass can become one entry covering shapes, shared filters and pooling. Two questions about padding stay out. The log records the scope you understood, so tracing a forward pass doesn't become a claim that you've learned neural networks.
Each entry has a topic, project, date and short explanation. It lives in a Markdown file you choose. Read it yourself, ask your agent about it, or use the separate Revise skill to return to a topic.
No scripts, API keys or extra services. Your agent must load the skill and have read/write access to the log. Capture depends on the model's judgment; it isn't a complete archive or a measure of retention.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Education & Learning.
Capture depends on your agent loading the skill and judging the available conversation. Brief exchanges and time spent alone don't qualify. Entries record observed understanding, not lasting retention. Existing history is not automatically consolidated.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free to use
No account required to browse
Trust & safety
Security scan
Security scan passed 3 days ago
- Free with an Agensi account
Nothing to install. It runs in your AI agent.
Choose your log file and ask the agent to use Learning Log while you work. No companion skill is required.
Permissions
File Scopes
Read/write access only to the log you choose. The default filename is learned-log.md; a custom path can be specified. No network calls or telemetry are bundled. Your agent's data policies still apply.