Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Architecture Opportunity Scan

    1

    Read-only scan that turns change friction, duplicated policy and knowledge hotspots in a codebase into a ranked table of structural opportunities, each with its cause, LEVERAGE, RISK, EFFORT and next design step.

    $16

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Architecture Opportunity Scan

    Architecture Opportunity Scan

    Example session with this skill installed

    Run an architecture opportunity scan on our Node.js order service. Keep it read-only and do not change any code. Git history is not available in this environment, so mark anything that needs history as UNVERIFIED.

    • Read your context and instructions
    • Compiled the architecture opportunity scan

    Illustrative worked example with synthetic data. File names, counts and findings are made up to show the shape of the output; in real use the skill reports only evidence supplied or verified in your environment.

    Runtime fallback: Git history and graph tooling absent -> historical evidence marked UNVERIFIED; used imports + duplicated policy as bounded evidence.
    
    OPPORTUNITY | EVIDENCE | CAUSE | LEVERAGE | RISK | EFFORT | NEXT DESIGN STEP
    Pricing rules duplicated in 3 modules | checkout.js, cart.js, invoice.js each re-implement tax rounding | missing single owner for pricing policy | HIGH | MEDIUM | LOW | route to architecture-seam-design: extract pricing module
    DB layer imported directly by 7 route files | grep of `import db` across routes | leaked infrastructure detail, no repository seam | HIGH | MEDIUM | MEDIUM | design repository boundary
    auth check copied per-route | 9 near-identical guards | duplicated policy | MEDIUM | LOW | LOW | middleware/ownership helper
    
    No implementation performed. Each opportunity is evidence-backed with a bounded next step.
    

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

    About this skill

    Architecture Opportunity Scan answers one question for a codebase that has become hard to change: which structural fixes are worth the migration cost? It looks for concrete change friction (the same edit repeated across modules, duplicated policy, unstable dependencies, broad interfaces, knowledge hotspots), traces each finding to an architectural cause such as a missing seam or wrong ownership, and returns a ranked table of evidence-backed opportunities. It is read-only: it reports, it does not refactor.

    What it does

    • Finds structural work worth doing and ignores aesthetic churn.
    • Establishes friction from evidence such as repeated multi-module edits, duplicated rules, circular or unstable dependencies, broad interfaces, hard-to-isolate tests, failure coupling and knowledge concentrated in surprising places.
    • Explains the architectural cause instead of restating a code smell, and distinguishes a missing seam, wrong ownership, leaked infrastructure detail, a stale boundary and accidental duplication.
    • Rates each opportunity LOW, MEDIUM or HIGH on LEVERAGE, RISK and EFFORT, with evidence, and rejects opportunities whose benefit is mainly stylistic or speculative.
    • Proposes the smallest structural move: the target seam or ownership change, migration steps, affected contracts, validation strategy and a stop condition.

    How it works

    1. Establish friction from the architecture docs, recent change patterns when available, module boundaries, imports and dependencies, repeated policy, tests and a bounded set of representative hotspots.
    2. Trace causes for each candidate.
    3. Score opportunities. Work is recommended only when the expected reduction in future change cost exceeds migration and regression risk.
    4. Propose the smallest structural move and name the next design step. Design-heavy candidates are routed to architecture-seam-design before any implementation.

    What you get

    A table with one row per opportunity:

    OPPORTUNITY | EVIDENCE | CAUSE | LEVERAGE | RISK | EFFORT | NEXT DESIGN STEP
    

    Every row is evidence-backed, carries a causal explanation and a bounded next step, and no implementation or broad refactor has been performed.

    Evidence discipline and boundaries

    • Read-only: it does not rewrite code, and dependency install, commit, push, merge, deploy and destructive migration are never implied.
    • It reads the smallest sufficient evidence (exact files, ranges, diffs, failing tests) instead of dumping the repository.
    • It never assumes a shell, Git, browser, network or MCP tool exists. If Git history, dependency graphs, metrics or code-index tooling are absent, it falls back to current imports, duplicated policy and module contracts, and marks unavailable historical or graph evidence UNVERIFIED.
    • Project governance documents, accepted ADRs, committed contracts and tests outrank the skill. It guides process; it does not silently override your project rules.
    • Not for a known local bug, automatic refactoring, an approved implementation or an aesthetics-only cleanup.

    How it differs from nearby work

    This skill finds and ranks structural opportunities across a meaningful codebase surface and stops at a bounded next design step. Designing the target seam for a chosen candidate is a separate step, which is where architecture-seam-design comes in.

    Use cases

    • Audit our architecture and rank where structural work pays off first.
    • Find the high-leverage cleanups in a codebase that has become hard to change.
    • Locate business rules that are re-implemented across modules.
    • Separate structural technical debt from cosmetic cleanup before committing time to a refactor.

    Validation

    The package includes one worked example, labeled as illustrative with synthetic data. No benchmark, measured outcome or host-validation result is claimed for this listing.

    Known limitations

    See the Known Limitations section of this listing.

    This is a discovery and ranking process, not an implementation or design tool: it does not refactor, does not produce the target seam design, and stops at a bounded next design step. Ratings are qualitative (LOW, MEDIUM, HIGH), not measured metrics. Findings are limited to the evidence available to the host; when history, dependency graph or metric tooling is absent, related findings are marked UNVERIFIED. The skill guides the agent's process and does not guarantee model output, and your project governance documents and ADRs outrank it. The package contains one synthetic worked example and no benchmark or host-validation results.

    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

    Permissions

    Read Files

    Read-only scan of the repository under review: the skill reads project files and does not write, commit, push, install dependencies or deploy. Shell access is optional and only relevant when your environment offers Git history, dependency graph, metric or code-index tooling; without it the scan relies on file reading alone. No network access, browser, MCP service or environment variables are required.

    Listed19 days ago

    What's inside

    Frequently Asked Questions

    It tells you which structural fixes in a hard-to-change codebase are worth doing. It ranks opportunities by change friction, duplicated policy, interface size and knowledge hotspots, explains the architectural cause of each, and rates LEVERAGE, RISK and EFFORT so you can skip aesthetic churn.

    Use it when you ask where to improve a codebase's structure, where the technical debt is, or what is worth refactoring, for example 'audit our architecture' or 'where should we refactor first'.

    This skill discovers and ranks candidates across a meaningful codebase surface and stops at a bounded next design step. Design-heavy candidates are routed to architecture-seam-design, which is where the target seam is designed.

    No. It is read-only discovery. It does not rewrite code, and dependency install, commit, push, merge, deploy and destructive migration are never implied.

    An agent host that supports SKILL.md skills and can read the repository you want scanned. The package ships no scripts and needs no network access. Git history, dependency graph, metric or code-index tooling is optional: without it the skill falls back to current imports, duplicated policy and module contracts, and marks unavailable historical or graph evidence UNVERIFIED instead of guessing.

    New in Code Quality & Review