Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Senior Architect

    1

    Staff-engineer design reviews that start from repo facts: health snapshot, decision records, and a playbook for AI agents.

    Free

    2 usesSecurity scanned
    Senior Architect

    Senior Architect

    Example session with this skill installed

    Review the architecture of this repo before we add more features. Where are the risks?

    • Read your context and instructions
    • Compiled the senior architect

    Facts first — scripts/repo_health.py (real run on a three.js/React app)

    MetricValue
    Code138 files, 14,682 lines (TypeScript 15,229, JavaScript 985)
    Tests1,532 lines — test/code ratio 0.10
    Largest filesGlobeView.tsx 451, App.tsx 439, orbitEngine.ts 401, ArSky.tsx 399
    Runtime import cycles0
    Most importedlib/satellites.ts 16×, globe/helpers.ts 16×, lib/quakes.ts 11×
    Duplicated blocks3 groups — e.g. the same GLSL noise function in starMaterial.ts:31 and sunMaterial.ts:24

    Review

    1. No god modules, no cycles — the structure is healthy; do not reorganise for its own sake.

    Test ratio 0.10 is the real risk. satellites.ts and helpers.ts are imported 16× each: a change there ripples everywhere. Cover those two first.
    3.

    Shader duplication (noise in two materials) — rule of three: extract a shared noise.glsl only if a third copy appears.
    4.

    One-way door ahead? If the next feature adds accounts or a backend, record it: scripts/adr_new.py "Add user accounts".

    Not doing: splitting the 400–450-line components — they are cohesive, and splitting would add indirection without reducing risk.

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

    About this skill

    Technical debt and architectural rot kill products faster than competitors do. This skill provides the senior-level judgment needed to structure systems that remain maintainable for years, moving beyond simple code generation to strategic software design. It enforces a pragmatic, veteran approach to refactoring, schema design, and agentic workflows.

    What it does

    • Repo health audits runs diagnostics to identify import cycles, code duplication, and god modules.
    • Architectural Decision Records documents irreversible "one-way door" choices using a structured ADR format.
    • Complexity sizing evaluates requirements to prevent premature abstraction or unnecessary microservices.
    • Agent system design applies specific failure-mode patterns and context-engineering rules to LLM pipelines.
    • Anti-pattern detection flags copy-paste drift, resume-driven development, and leaky abstractions.

    How it works

    1. Analyze the current codebase facts using the repo_health.py tool to establish a baseline of technical debt.
    2. Review proposed changes against the veteran playbook, focusing on maintenance cost and failure modes.
    3. Execute architectural decisions by generating a numbered ADR that outlines context, options, and consequences.
    4. Refactor complex modules into vertical slices or isolated boundaries based on the rule of three.

    Frameworks & tools

    This skill includes custom Python scripts for repository analysis (repo_health.py) and ADR management (adr_new.py). It contains specific design wisdom for React state management, multi-tenant SaaS architectures, Supabase/PocketBase security, and MCP server implementations.

    Why this beats prompting it yourself

    Generic prompts often result in "architecture astronaut" advice that is too academic for real-world shipping. This skill enforces a deterministic-first approach, quoting hard repo metrics before offering opinions and recording the "why" behind decisions so you do not repeat past mistakes.

    Use cases

    • Splitting a monolithic file or god class into maintainable modules.
    • Deciding between a single agent, multiple agents, or a simple script for a task.
    • Designing a multi-tenant database schema with Row Level Security.
    • Isolating legacy ERP or payment integrations behind anti-corruption layers.

    Known limitations

    The repo health tool requires a Python 3 environment and access to the local filesystem. It focuses on structural patterns rather than language-specific syntax errors or linting.

    Requires Python 3 for health scripts. Analysis is limited to structural health and design patterns, not syntax linting or runtime debugging.

    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

    2 uses

    Times used to date

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 1 month ago

    • Free with an Agensi account

    Requires

    Python 3.9+
    Python

    Permissions

    Terminal / Shell
    Read Files
    Write Files

    Terminal runs the bundled read-only tools (repo_health.py, adr_new.py). Read Files: your repository. Write Files: adr_new.py creates one decision record in docs/adr/. No network access.

    Listed1 month ago

    What's inside

    Frequently Asked Questions

    The Senior Architect skill acts as a strategic layer for your development agent, using custom scripts to audit codebase health, document ADRs, and prevent 'architecture astronaut' patterns that lead to technical debt.

    The skill includes a suite of custom Python tools (`repo_health.py` and `adr_new.py`) along with a veteran design playbook that provides specific guidance for React, SaaS, and agentic workflows.

    The repo health diagnostics require a Python 3 environment and filesystem access; it is designed to work with any agentic framework that supports local tool execution and context engineering.

    While it provides high-level architectural guidance for most stacks, it contains specialized, optimized design patterns for React state management, Supabase/PocketBase security, and MCP server implementations.

    No, this skill is designed to prevent 'resume-driven development' by using a deterministic-first approach, prioritizing simple scripts and vertical slices over unnecessary microservices or complex agent swarms.

    New in Code Quality & Review