More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    GuideBranch — Support SOPs to Interactive Guides

    1

    Turn approved support SOPs into a source-linked troubleshooting guide with explicit unknown routes and an escalation summary. Includes a standalone builder; no API subscription.

    $12

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    GuideBranch — Support SOPs to Interactive Guides

    GuideBranch — Support SOPs to Interactive Guides

    Example session with this skill installed

    Fictional local demonstration: use the included TableNest CSV export SOP to create a closed-choice client guide, with source-linked nodes and a Not sure escalation route. Preserve the original instructions and return the HTML, flow data and author reviews.

    • Read your context and instructions
    • Compiled the guidebranch
    • Generated the data export

    Delivered artifacts

    Fictional TableNest demonstration, built locally from the included approved example SOP. The actual reader is shown in the screenshots.

    Release 0.1.1 reproduced the same HTML in an isolated temporary-helper build. The browser paths and summary-download checks below record the original, unchanged reader.

    FileDelivery
    guide.htmlReader guide with questions, Back/Start again and local summary downloads
    flow.jsonAuthored project data for future revisions
    review.jsonAuthor-only source hashes, exact citations and all enumerated paths
    review.mdAuthor-only readable review and limits

    Complete builder result

    {
      "status": "built",
      "flow_id": "tablenest-csv-export",
      "nodes": 10,
      "edges": 12,
      "sources": 1,
      "paths": 9,
      "source_hashes": {
        "tablenest-csv": "16f83a8e7315f5af9902754865d49142abc096082153803291057b03ed8b70fe"
      },
      "source_approval_verified": false,
      "semantic_correctness_verified": false,
      "reader_content_review_required": true,
      "output_dir": ".local/guidebranch-20261102/browser/tablenest",
      "files": [
        "review.json",
        "review.md",
        "flow.json",
        "guide.html"
      ]
    }
    

    Every TableNest path exercised in the browser

    AnswersReached guidanceType
    A file appeared → Yes, it opens → The headings matchReview the downloaded rowsnext_step
    A file appeared → Yes, it opens → The headings differConfirm the intended saved reportstop
    A file appeared → Yes, it opens → Not sureAsk the support ownerescalate
    A file appeared → No, it does not openAsk for import instructionsescalate
    A file appeared → Not sureAsk the support ownerescalate
    No file appeared → Yes, Export queued is displayedWait for the existing exportnext_step
    No file appeared → No queued banner is displayedTry Export CSV once morenext_step
    No file appeared → Not sureAsk the support ownerescalate
    Not sureAsk the support ownerescalate

    Back returned to the observed parent at every traversed edge, and route changes focused the heading. A second supplied example had 12 exercised paths; an independently authored flow had 15. Actual Markdown/JSON downloads contained only the current answer path after an earlier answer changed.

    Owner handoff

    Review the source meaning and each remedy before operational use. The helper has not verified source approval or semantic correctness. Keep exact quotations in the author review files separate from the reader handoff.

    Browser qualification used localhost. Direct file-URL opening was not tested because the browser tool permits only HTTP/HTTPS. This is a fictional maintainer demonstration, not an Agensi customer-use, ticket-reduction or income claim.

    guide.html

    CSV · data export

    Generated

    Example file from a real run - the skill writes it into your workspace.

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

    About this skill

    Give your agent the software-support instructions your team already approves, and ask for a guide people can follow one question at a time. GuideBranch turns the supplied procedure into a working, self-contained HTML guide and a separate author review pack.

    Try this request

    Use GuideBranch on this approved support runbook. Make a guide for our client team, preserve each remedy and escalation rule, include a Not sure route, and return the HTML with source-linked review notes.

    What you get

    • A polished troubleshooting guide with closed choices, Back, Start again and a readable answer trail.
    • Locally downloadable Markdown and JSON summaries for a support handoff. Nothing is sent automatically.
    • An authored flow file for later updates, plus exact source quotations, source hashes and enumerated paths in separate author-only review files.
    • Two included fictional examples for understanding the workflow.

    The agent does the editorial work and prepares the flow data. The included Python builder checks exact citations, missing targets, duplicate IDs, unreachable nodes, cycles, terminal outcomes and explicit unknown-answer routes. Missing or conflicting remedies stay unresolved; a structural pass does not certify the source policy or its meaning.

    Before you start

    Use an Agensi-connected agent with Python 3.10+, local file/shell tools and a browser preview. Approved UTF-8 TXT/Markdown sources are supported directly. The independent helper runs unchanged from one temporary file, without pip, another MCP server, an API key or separately metered processing.

    Delivery and boundaries

    The HTML includes all its branches and has no access control. Keep private staff instructions out of reader text, and keep the author review files separate. The agent and execution host still have their usual access to supplied material. Hosting and customer distribution are separate decisions; a local file is not an automatically hosted client URL.

    For ordinary software and digital-product support. No automatic system actions, ticket submission, analytics or secret collection. Not for medical, legal, financial, emergency or safety-critical decision guides. Maximum 100 nodes, 400 edges, 1,000 paths and 20 sources, with 2 MiB per source/flow.

    Tested with fictional support procedures, an independently authored guide, and complete paths through generated browser guides. Release 0.1.1 also refuses a source replaced by a named pipe during opening on the tested POSIX host. These checks do not establish customer adoption, ticket reduction or guaranteed resolution. Browser checks used localhost; direct file opening was not verified in this session.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Technical Documentation.

    Ordinary software and digital-product support only. Source TXT/Markdown up to 2 MiB each; 100 nodes, 400 edges, 1,000 paths, 20 sources. Exact citation and graph checks do not verify source approval, policy correctness or semantic fidelity. All authored branches are inspectable in the HTML; source quotations stay in separate author review files. No hosting, access control, automatic action or ticket sending. Browser verification needs an available authorized preview. No medical, legal, financial, emergency or safety-critical decision guides.

    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 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

    Money-back guarantee

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

    Trust & safety

    Security scan

    Security scan passed 15 days ago

    Software
    Python 3.10+Terminal access
    Modern browser
    orbrowser preview

    The browser is used to verify interactions.

    Permissions

    Terminal / Shell
    Read Files
    Write Files
    Browser

    Read only selected approved source/flow files. Create a new output directory; preserve source files and previous releases. Run the unchanged helper temporarily. No environment-variable access, network calls, accounts or credential collection by the helper. Reader answers stay in memory; summaries download locally. All guide branches are inspectable. Author review files may contain private source quotations. Hosting, sharing and support messages require their own authorized scope.

    Listed15 days ago

    What's inside

    Frequently Asked Questions

    Approved TXT/Markdown support instructions, the intended audience and the support-owner role. The agent prepares the flow data and exact citations. You do not need to write JSON or configure another server.

    The one-time purchase includes the skill instructions, standalone Python builder and fictional examples. Your agent needs local file/shell tools, Python 3.10+ and a browser preview. There is no separate processing API subscription or key.

    Yes, within your authorized project scope. The generated HTML and its necessary inline interactivity are the work product. Keep the paid skill/helper source and author-private review files out of the client handoff. The HTML contains every branch and has no login; hosting is not included.

    No. It checks exact quotations and graph structure. It cannot prove a paraphrase is faithful, resolve contradictory policies or certify a remedy. The agent exposes those gaps for the source owner instead of inventing an instruction.

    It keeps closed-choice answers in browser memory and downloads Markdown/JSON summaries locally. It sends nothing automatically, uses no cookies or local storage, and runs no support action. Reloading clears the session.

    Missing or conflicting instructions produce cited author questions. An incomplete operational guide is not presented as approved. For updates, load the skill fresh, use revised sources and a new output directory/version, then repeat editorial and browser checks.

    New in Technical Documentation