Revise - Learning Log Review

    1

    Return to a topic in your learning log. Work through an explanation, answer a reasoning question, then go deeper or move on.

    Free

    Security scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Revise - Learning Log Review

    Example session with this skill installed

    Revise my CNN forward-pass reasoning. Teach first, then ask one question.

    My log has entries on convolution output shape, shared filters and pooling.

    • Read your context and instructions
    • Compiled the revise - learning

    Illustrative review based on a synthetic QA session with a four-entry test log. The wording is edited for clarity; the response stops for the learner's answer.

    A CNN forward pass builds maps of where patterns appear, then can shrink those maps while keeping their channels separate.

    Imagine an 8 x 8 image with 3 input channels. A convolution uses 4 filters, each covering a 3 x 3 spatial patch across the input channels. Each filter reuses the same learned weights at every position. The four filters produce four output feature maps.

    With stride 1 and no padding, a 3-wide filter fits in 6 positions along each side: 8 - 3 + 1 = 6. The convolution output is 6 x 6 x 4. Padding adds space around the input; a larger stride skips more positions.

    Next, 2 x 2 max-pooling with stride 2 keeps the largest value in each non-overlapping block, separately within each feature map. Each 6 x 6 map becomes 3 x 3, but there are still four maps. The result is 3 x 3 x 4.

    What output shape would you predict after the same convolution and pooling for a 10 x 10 x 3 input, accounting for both spatial size and channel count?

    Connects securely to your tools. The creator never sees your data.

    About this skill

    "Revise what I learned about CNNs last month."

    Revise reads your learning log and gives you a place to start. It explains one connected topic, then asks a question that requires you to use it. Your answer guides what needs another explanation.

    An older log might have separate entries for convolution shapes, shared filters and pooling. Revise can bring those together into a forward-pass review without editing the file or pretending you studied the whole field.

    Stay with a difficult point, ask for another example or move to a different topic. Teaching comes first unless you ask for active recall. If you don't choose a topic, it offers up to five recent ones.

    The log stays unchanged. There are no scores, streaks or extra review files to maintain. A compatible Markdown log is enough; Learning Log is optional. Bring the original conversation when you have it, since a one-line entry can't reproduce the exact lesson.

    Requires an agent that loads SKILL.md instructions and can read the log. No scripts, API keys or extra services.

    Requires a Markdown log in the documented format and an agent that can read it. A short log entry cannot reconstruct the original lesson or establish mastery. Feedback is limited to the current answer; retention isn't measured. No scheduled reviews or review-history files.

    How to use

    1. 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. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    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 forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 3 days ago

    • Free with an Agensi account

    Requires

    A compatible Markdown learning log

    Permissions

    Read Files

    File Scopes

    **/learned-log.md

    Reads your chosen learning log without changing it. Can also read original notes you provide for context. No network calls, telemetry, scores or review-history files are bundled.

    Listed3 days ago

    What's inside

    Frequently Asked Questions

    No. Any log in the supported Markdown format works: topic, project, date and short explanation.

    No. Related entries can be grouped for the conversation while the source file remains unchanged.

    Yes. Ask for active recall. Otherwise it teaches first, then asks one reasoning question.

    No. Feedback reflects your answer in that session, not a measurement of lasting retention.

    Yes. Specify a topic, project or date range. Without a specific request, Revise offers up to five recent topics. You can stay with a difficult detail or move on.

    The free ZIP contains SKILL.md instructions, a README with setup and examples, and the MIT license. It bundles no scripts, API keys or external services.

    New in Education & Learning