Table GitHub Selection Catalogue

    1

    Compare GitHub tools with cited evidence, explicit unknowns and weighted criteria. Keep a reusable, searchable capability catalogue without installing candidate repositories.

    $19

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Table GitHub Selection Catalogue

    Example session with this skill installed

    Compare these two GitHub repository candidates for local CSV analysis. Offline processing and a permissive licence are required. Prefer a documented JSON export. Keep unknowns visible, compare against using the Python standard library, and save a catalogue I can search later. Do not install either project.

    • Read your context and instructions
    • Compiled the table github selection

    Fictional worked example—not real repository claims.

    Candidate A ranks as eligible for further consideration: 75–100 weighted score, 75% usable-evidence coverage. Its export feature is only a README claim, so the next step is source verification. Candidate B remains provisional because offline processing lacks the required evidence. The report includes the no-extra-tool baseline, source citations and dates, the full criterion breakdown and the next check for each candidate. Neither project is installed or declared secure.

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

    About this skill

    A good GitHub shortlist needs more than stars and confident prose. Table GitHub Selection Catalogue helps your AI agent turn repository candidates into structured decision records you can revisit.

    You choose the task, hard requirements and weights. Your agent collects evidence through its authorized browsing or GitHub tools. The compact offline Python helper normalizes repository identity, validates records, merges nonconflicting candidates, compares evidence against your criteria and searches saved capabilities.

    Instead of hiding missing information, the comparison shows lower/upper scores, usable-evidence coverage, unmet requirements and the next fact to verify. README claims stay separate from source checks and real test records. A strong provisional candidate is a research lead—not automatic approval to install.

    Included:

    • A focused collection → catalogue → comparison → retrieval workflow.
    • An original 13 KB, dependency-free Python helper with ingest, compare and query commands.
    • Stable URL-hash IDs that preserve distinct repository names; exact duplicate handling.
    • Atomic new snapshots that refuse to overwrite existing output.
    • Criterion-level evidence dates, citations, status, freshness rules and explicit unknowns.
    • A fictional worked example with outputs generated by the included helper.
    • Source attribution and file-level licence notices.

    The broader Table-GitHub-Capability-Router is free and MIT-licensed. This paid package adds a new self-contained helper and a narrower, delivery-ready selection workflow. It is not an exclusive copy of the public repository or an automatic installation manager.

    Requirements

    An agent able to read delivered text assets, collect repository evidence using authorized tools, write and execute a temporary Python script, and retain your output data. Existing Python 3.9 or newer; local writable data directory; filesystem hard-link support for atomic output creation. No package installation, API key or paid external service is required by the offline helper. Existing browsing/GitHub access may depend on your agent account.

    Only Linux has been tested. Other operating systems and actual Agensi buyer delivery are unverified. Code uses Python 3.9-compatible syntax; runtime tests in this release used Python 3.12. No universal compatibility promise.

    Limits

    The helper does not browse, verify citations, execute candidate repositories, audit security or benchmark performance. Your agent supplies evidence and must assess its accuracy. It does not implement semantic search, background sync, hooks or persistent agent configuration. Each input/output is limited to 2 MB and catalogues to 1,000 records. Updating evidence requires a reviewed new snapshot; no silent conflict resolution. User data remains persistent; fetched reusable runtime must be removed after each use.

    The helper does not browse, verify citations, execute candidate repositories, audit security or benchmark performance. Your agent supplies evidence and must assess its accuracy. It does not implement semantic search, background sync, hooks or persistent agent configuration. Each input/output is limited to 2 MB and catalogues to 1,000 records. Updating evidence requires a reviewed new snapshot; no silent conflict resolution. User data remains persistent; fetched reusable runtime must be removed after each use. Only Linux has been tested; runtime tests used Python 3.12. Other systems and Agensi buyer delivery remain unverified.

    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 19 days ago

    Requires

    GitHub
    Python 3.9+
    Terminal access
    Writable data directory
    Web browsing or GitHub tools
    Python

    Permissions

    Terminal / Shell
    Read Files
    Network Access
    Write Files

    Allowed Hosts

    github.com

    Reads input records and writes user-authorized catalogue snapshots and reports. Runs the temporary Python helper, then removes that reusable runtime. Network access refers only to the agent collecting GitHub evidence through its existing authorized tools; the helper itself is offline. No candidate installation, candidate execution, mandatory API key or account connection.

    Listed19 days ago

    What's inside

    Frequently Asked Questions

    No. It evaluates and records candidates. Running or installing a candidate requires separate authorization and is outside the helper.

    No. The helper validates structure and calculates transparent rankings. Your agent collects and evaluates evidence; supplied citations and test records are not independently authenticated by the program.

    The free project is a broader repository-based routing workflow. This package has an independently authored compact catalogue/comparison core, immutable snapshots and weighted uncertainty-aware comparisons without that repository's dependencies.

    Yes. Your data and generated reports are persistent outputs. The reusable runtime is temporary and is removed after execution, consistent with platform delivery rules.

    Existing Python is required for the helper; it makes no network calls. Repository fact collection uses whatever authorized browsing/GitHub access your agent already has.

    A task-specific weighted decision aid, not a probability, safety rating or benchmark. Unknown or insufficient evidence widens the score interval; unmet hard requirements exclude a candidate.

    Adapted upstream guidance remains MIT with its original notice. Newly authored helper and example assets are supplied under the applicable Agensi purchase licence. See the package provenance notice; no exclusivity is claimed over the free source.

    New in Research & Analysis