Works with the AI tools you already use
Claude Code
Cursor
Codex CLI
Muse
OpenClaw+21 moreClaude CodeCursorCodex CLIMuseOpenClaw
Manus
Grok
Grok Bot+18 agentsArchitectureAdvisor: Pattern Recommender
Detect architectural pain points and get specific pattern recommendations to fix messy codebases.
Free
ArchitectureAdvisor: Pattern Recommender
Example session with this skill installed
My main.py is getting huge and the agent is starting to lose track of functions. I'm also seeing some weird import errors. What am I doing wrong here?
- Read your context and instructions
- Compiled the architectureadvisor
I've scanned your project structure and identified three architectural bottlenecks that are likely causing your agent to hallucinate and your imports to fail.
Architectural Assessment
SIGNAL: Single file over 800 lines with mixed responsibilities.
PATTERN: Module split.
SHAPE: Extract logic into a directory structure: models/, services/, and handlers/. The main.py should only serve as the entry point to wire these modules together.
WHY IT MATTERS: AI agents typically lose track of file context around 1500 lines. Large files cause the agent to drop functions or reinvent helpers, turning the file itself into a bottleneck for development.
SIGNAL: Global connection or bot instance imported in leaf modules.
PATTERN: Dependency Injection.
SHAPE
# Instead of: from main import bot
# Use:
def handle_update(bot, update):
bot.send_message(update.chat.id, "Shared instance")
WHY IT MATTERS: Importing the main app instance into sub-modules creates circular dependencies. This leads to silent freezes at startup when the import order shifts during refactoring.
SIGNAL: Direct os.getenv() calls scattered across multiple files.
PATTERN: Config Broker.
SHAPE: Create a config.py that reads all environment variables into a typed class. All other modules must import from this central broker.
WHY IT MATTERS: When a secret needs rotation or a variable name changes, you have to hunt through the entire codebase. A central broker makes configuration predictable and manageable.
Next steps
- Decide which boundary to draw first, starting with the Module Split to help your AI agent regain context.
- Create a
config.pyto centralize your environment variables. - Pass your database or bot instances as arguments to functions instead of importing them from your entry file.
Connects securely to your tools. The creator never sees your data.
About this skill
Your codebase feels "off" but you cannot put a name to the mess. You are seeing circular imports, silent production crashes, or an AI agent that starts hallucinating as your files grow too large. This skill identifies the technical debt slowing you down and names the specific architectural patterns needed to fix it.
What it does
- Detects anti-patterns by scanning open files for specific signals like manual import rows or raw dictionary access at boundaries.
- Identifies missing resilience such as bare API calls lacking timeouts or fallback layers.
- Diagnoses scaling bottlenecks by flagging monolithic files that exceed the context window of most AI coding assistants.
- Recommends specific patterns like Router Factories, Config Brokers, and Dependency Injection without the fluff.
- Visualizes the fix by providing a short code fragment or structural description of the proposed solution.
How it works
- Trigger a scan by describing a symptom like "the agent keeps breaking this file" or "should I refactor this?".
- The skill analyzes your current file and project structure for eight specific architectural signals.
- You receive a report containing up to three detected signals, the corresponding pattern name, and the consequences of ignoring it.
Frameworks & tools
This skill is language-agnostic but excels in Python, JavaScript, and TypeScript environments where decorators, environment variables, and modular imports are common.
Why this beats prompting it yourself
Generic prompts often result in lengthy lectures on SOLID principles or suggestions for complex libraries you do not need. This skill focuses on the "vibe-coder" pain point, translating vague discomfort into concrete, industry-standard architectural names that help you and your agent work better together.
Use cases
- Identify why your AI agent is losing track of functions in large files.
- Secure environment variables by moving from scattered
os.getenvcalls to a Config Broker. - Stabilize external integrations by adding resilience wrappers to bare network calls.
- Prevent circular dependency freezes by identifying where to apply Dependency Injection.
Known limitations
This is an advisory tool only. It will not implement the code changes for you. It ignores projects under 300 lines where these patterns are often overkill.
How to install
Works the same in every agent - Claude, Cursor, Codex, Copilot and 20+ more.
- 1
Download the ZIP
Free skills download straight away. Paid skills unlock right after purchase.
- 2
Unzip into your skills folder
Every agent reads skills from one folder on your machine. Drop the unzipped folder in there.
- 3
Ask your agent to use it
Restart the agent if it was already running. It picks the skill up automatically - no config needed.
Skills folder by agent
Click the path to copy it. Create the folder if it does not exist yet.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill passes our trust checks below.
Security scanned
Passed our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free forever
No account required to browse
Trust & safety
Security scanned
Verified clean 1 day ago
- Free to download with an account