Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Context Library Keeper Pro

    1

    **Inventory** — lists every item in the library with a one-line summary written from its actual content, plus source, date, and flags, in a single table.

    $9.99

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned

    About this skill

    Context Library Keeper Pro turns a knowledge library that outgrew its owner — hundreds of notes, notebooks, and saved references across folders and accounts — into one index you can actually scan.

    What it does

    1. Inventory — lists every item in the library with a one-line summary written from its actual content, plus source, date, and flags, in a single table.
    2. Count reconciliation — breaks the total down (by date, folder, or account) and proves the parts sum to the whole before reporting. A report that doesn't reconcile isn't finished.
    3. Multi-store merge — libraries split across two accounts, a drive, and a notebook app are reconciled into one master index, with each item's home store kept.
    4. Flag engine — duplicates (with evidence, not just similar titles), clusters of related items, empty stubs, items superseded by newer coverage, and personal-information items marked "handle with care".
    5. Supersession tracking — when a newer note replaces an older one, the index links them. The old note stays as the source record; replacement is a relationship, not a deletion.
    6. Scheduled upkeep — recurring runs report only the diff: new items, new flags, new supersessions, and the public arithmetic (previous + new − removed = current).
    7. Approval-gated pruning — candidates are presented first; nothing is deleted or moved before you approve the list; every action lands in a restore log.

    The safety design (the point of this skill)

    Cleanup automations that delete on their own are how people lose the one note they needed. This skill is built the opposite way:

    • Approval gate — it presents candidates and waits. Deleting, moving, merging, and renaming all require your explicit OK on the specific list.
    • Personal means protected — anything that looks like personal information is flagged "handle with care" and excluded from every pruning list, regardless of age or size.
    • Read, not guessed — summaries come from opening the item. Anything it couldn't inspect is labeled "not inspected", never summarized from its title.
    • Zero is a result — if your library has no duplicates and no empty stubs, the report says so as a health check, with the reconciled counts as proof.

    See it in action

    User request (sample run, fictional data): "My research notes folder has gotten out of hand. Inventory it and tell me what needs attention."

    What the skill produced (excerpt):

    • 132 items inventoried (41 + 63 + 28 by month = 132, reconciled)
    • 3 duplicate groups flagged with evidence — in each case the same saved article filed twice under different names; one canonical copy recommended per group
    • 11 items clustered as one fast-moving topic covered by multiple sources, cross-referenced in the index instead of merged
    • 2 empty stubs (title only, no content) listed as pruning candidates; 1 contract-like document flagged "handle with care" and excluded from candidates
    • 4 older notes linked as superseded by newer deep-dives, kept as source records
    • Nothing deleted: the candidate list was presented and left for the user's decision

    • It can only inventory what it can read: libraries behind access you don't grant stay uninspected and are listed as such.
    • Judging whether a superseded note still matters is the user's call — the skill links and explains, it doesn't decide.
    • Pruning is executed only where the platform gives the agent a tool path; otherwise it produces the exact steps for the user.

    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

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 6 days ago

    Permissions

    No special permissions declared or detected

    Listed6 days ago

    What's inside

    Frequently Asked Questions

    Never. It presents a candidate list and waits for your approval of that specific list. Items that look like personal information are flagged 'handle with care' and excluded from every pruning list.

    Any library your agent can read: note folders, connected drives, notebook apps it has access to. Anything it cannot open is listed as 'not inspected' and is never summarized from its title alone.

    It links the older item to the newer coverage as 'superseded', shows the evidence, and leaves the keep-or-prune decision to you. Superseded notes stay as source records by default.

    Every report reconciles: the breakdown must sum to the total before the report is finished, and recurring runs show previous + new - removed = current in the open. A mismatch is reported, never smoothed over.

    New in Productivity