Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    Architect Docs Engine Premium

    1

    Turn a rough idea or existing codebase into a reviewable engineering handoff: 21 connected architecture docs plus four practical files for decisions, delivery, security verification, and UI review.

    $5

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Architect Docs Engine Premium

    Architect Docs Engine Premium

    Example session with this skill installed

    I'm starting a bicycle repair booking app from zero. Customers request a service and preferred time; staff confirm appointments and update repair status. I have no repository, tech stack, database, hosting, or login design yet. Create the 21-document suite and all four Premium handoff files. Keep unknowns marked Unresolved and recommendations marked Proposed; do not invent decisions. Trace requirements to decisions, implementation tasks, and verification checks. Focus security on booking privacy, unauthorized access, duplicate submissions, and booking spam; write safe local checks only. Keep the main booking form on a page or inline, not in a popup. Save files under docs/ and planning/architect-docs-engine/. Do not write application code.

    • Read your context and instructions
    • Compiled the architect docs engine

    Sample handoff: bicycle repair booking

    Starting evidence: a short product idea only. No repository, stack, database, hosting, login policy, screen designs, or expected traffic was provided. Missing choices remain

    Unresolved; recommendations are

    Proposed. No code was written and no tests were run.

    Complete file set

    The handoff contains 21 canonical files in docs/: README, PRD, TECH_STACK, DECISIONS, DATA_MODEL, VALIDATION_RULES, API_CONTRACT, UI, UX, AGENTS, PROMPTS, SECURITY, SEED, IMPLEMENTATION, TESTING, ACCEPTANCE, TRACEABILITY, OPERATIONS, AUDIT, _AUDIT_REPORT, and _CHANGELOG.

    Premium adds four linked files under planning/architect-docs-engine/: DECISION_REGISTER.md, IMPLEMENTATION_BACKLOG.md, SECURITY_VERIFICATION_MATRIX.md, and UI_REVIEW_SCORECARD.md. The 21 canonical files remain in docs/; the companion files do not replace or duplicate them.

    Decision register

    IDStatusDecision and evidence needed
    ADR-001UnresolvedCustomer identity is unspecified. Compare customer accounts with private booking links against privacy, recovery, and ease of use before approval.
    ADR-002UnresolvedThe stack is unknown. Choose after hosting, maintenance skills, and expected booking volume are understood.
    ADR-003ProposedStaff should confirm a requested appointment before treating it as booked. Confirm this matches the shop's real workflow.

    Implementation backlog

    IDPriorityTaskAcceptance and verification
    TASK-001P0Define who may view, change, and cancel a booking.Document the identity and ownership rule; get an owner to approve it before implementation.
    TASK-002P0Validate booking details and appointment times on the server.Invalid or unavailable requests are rejected with useful field guidance; a valid request creates one pending booking. Test with synthetic data in a local environment.
    TASK-003P1Prevent accidental duplicate requests and automated booking abuse.Repeated submissions do not create unintended duplicate jobs; apply a server-side rate/resource limit based on expected use. Verify locally.
    TASK-004P1Build a staff view of pending requests and repair status.Staff can find and update only bookings they are authorized to manage. Verify using separate synthetic customer and staff accounts.

    Security verification matrix

    ThreatPlanned controlSafe checkStatus
    THREAT-001: one customer reads another customer's bookingServer-side ownership authorization on every read and updateIn an isolated local test, sign in as customer A and request customer B's synthetic booking ID; expect denial and no booking data in the response.Proposed; Not run
    THREAT-002: repeated or automated booking requests overwhelm staffServer-side duplicate protection, rate limit, and request-size/resource boundsSend a small, bounded set of requests to a local test instance; confirm the documented limit and normal recovery. Never target a live service.Proposed; Not run
    THREAT-003: unauthorized staff changes a bookingRole checks on each staff actionTry a staff-only update as a customer test account; expect denial and an audit-safe event.Proposed; Not run

    UI review scorecard

    FLOW-001 — Request a repair slot. Proposed layout: one page or inline form for service, preferred time, contact details, and a clear submission status. A main-form popup is not justified by this workflow. Show field-level errors and distinguish “request received” from “appointment confirmed.”

    Not checked: no screen or prototype was supplied, so color, typography, spacing, icon meaning, responsive layout, keyboard flow, focus behavior, and accessibility cannot yet be verified. Review at mobile and desktop sizes when a real screen exists.

    Handoff and next steps

    1. Confirm how customers identify themselves and how appointment slots are approved.
    2. Decide stack and hosting after expected shop volume and maintenance needs are known.
    3. Approve the booking privacy rule before implementation.
    4. When a prototype or local implementation exists, replace Proposed/Not run with evidence from actual review and safe checks.

    This is a planning handoff, not proof that controls have been implemented or that the application is secure.

    DECISION_REGISTER.md, IMPLEMENTATION_BACKLOG.md

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

    About this skill

    Turn “we should build this” into an engineering handoff your team can actually review.

    Architect Docs Engine Premium starts with either a rough product idea or an existing codebase. It creates the canonical 21-document architecture suite, then adds four separate working files to carry the plan into decisions, implementation, security review, and interface review.

    What you receive

    • 21 connected Markdown documents in docs/, covering product requirements, technical choices, data and API contracts, UI and UX, security, implementation, testing, acceptance, traceability, operations, and audit history.
    • Decision register: compare options using evidence and trade-offs; mark approvals, owners, and revisit conditions so open choices do not disappear into chat.
    • Implementation backlog: turn requirements into ordered, bite-sized tasks with dependencies, acceptance checks, negative cases, and verification evidence.
    • Security verification matrix: connect product-specific threats to controls and safe, repeatable checks. Authorization, abuse and bot limits, resource bounds, data handling, and agent risks are included when relevant.
    • UI review scorecard: check real user tasks, layout, hierarchy, color, typography, spacing, icons, responsive states, and accessibility evidence. Main forms default to page-level or inline layouts; a modal needs a product-specific reason.

    The four companion files live in planning/architect-docs-engine/ and link back to docs/. That traceability helps a team see why a requirement exists, what work follows, and what evidence would show the work is done.

    What makes the handoff dependable

    The skill labels inspected facts, user-provided details, proposals, unresolved decisions, and checks that have not been run. Missing information stays visible instead of being invented. A short brief is enough to start; for an existing project, the agent can inspect relevant files as evidence.

    The free GitHub edition already includes the 21-document workflow and local inspection and validation tools. Premium adds four separate team-work artifacts; it does not claim the free edition lacks UI or security guidance.

    Requirements and limits

    Use a compatible AI coding agent with read/write access to the project workspace. No API key, CLI, or extra package is required. This skill creates documentation only: it does not write application code, execute project code, perform live penetration tests, or certify security. Proposed controls remain unverified until the project team implements and checks them.

    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.

    Creates documentation only; it does not implement application code, execute project code, test live systems, or certify controls. Security checks remain plans until safely run against an authorized test environment. UI findings require real screens or prototypes for a verified review.

    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 7 days ago

    Nothing to install. It runs in your AI agent.

    For an idea-only start, provide a short product brief. For an existing project, the agent reads relevant project files and writes documentation only under docs/ and planning/architect-docs-engine/.

    Permissions

    Read Files
    Write Files

    File Scopes

    **/*

    Read access covers relevant project evidence; write access is for docs/ and planning/architect-docs-engine/ only. No shell, network, or external service access is needed.

    Listed7 days ago

    What's inside

    Frequently Asked Questions

    Both editions include the 21-document architecture workflow. Premium adds four linked working files: a decision register, implementation backlog, security verification matrix, and UI review scorecard. They make choices, next steps, safe checks, and interface findings easier for a team to review and track.

    Any compatible agent that supports the standard SKILL.md format and can read and write the selected project workspace. The package has no CLI or runtime dependencies. It needs file reading and writing only.

    Yes. The agent inspects relevant source, configuration, tests, and existing docs as read-only evidence. It does not execute project code. Existing authored files are preserved; the agent proposes a scoped merge before replacing meaningful content.

    No. It creates architecture and verification documentation; it does not implement application code, run scans against live systems, or certify security. Checks stay Not run until evidence shows they were performed.

    This submission is version 1.1.0. There is no promised update schedule. If the creator publishes a later version, check the listing for its release details.

    New in Technical Documentation