chatgpt canvas workflow pro
A disciplined multi-pass workflow for precise document editing and code refactoring inside ChatGPT Canvas.
$5
Works with the AI Agents you already use
Claude Code
Cursor
Codex CLI
Muse
Dots+21 moreClaude CodeCursorCodex CLIMuseDots
Manus
Grok
Grok Bot+18 agentschatgpt canvas workflow pro
Example session with this skill installed
I have a Python function in Canvas that is failing silently. I need you to run a debug pass to find the bug and then add logging to every early return.
- Read your context and instructions
- Compiled the chatgpt canvas workflow
I will start the debug pass on your Python function. I'll first add explanatory comments to identify the logic flow, then locate the silent failure.
Debug and Logging Audit
| Step | Action | Description |
|---|---|---|
| 1 | Logical Mapping | Inserted # comments above lines 12, 18, and 24 to trace variable state. |
| 2 | Bug Detection | Identified silent failure at line 29: the except block catches Exception but does not log or re-raise. |
| 3 | Logging Insertion | Added logger.debug() calls to the three early return paths at lines 15, 21, and 34. |
| 4 | Fix Proposal | Proposed changing pass to logger.exception("Validation failed") in the catch block. |
import logging
logger = logging.getLogger(__name__)
def validate_user_session(session_id, user_data):
# Check if session_id is provided
if not session_id:
logger.debug("Early return: Missing session_id")
return False
# Verify user_data contains required keys
if "uid" not in user_data:
logger.debug("Early return: Missing uid in user_data")
return False
try:
# Simulate database lookup that might fail
record = db.fetch(session_id)
if not record:
logger.debug("Early return: No record found for session_id")
return False
except Exception:
# BUG: This was previously a silent 'pass'
logger.exception("Database connection failure during session validation")
return False
return True
Next steps
- Paste the updated code into your Canvas document.
- Verify that your logging configuration (level/handler) is set to capture
DEBUGoutputs. - Confirm if you want to replace the generic
Exceptioncatch with a specificDatabaseError.
Connects securely to your tools. The creator never sees your data.
About this skill
The problem
Large-scale editing and coding in a standard chat window lead to context loss, truncated code blocks, and messy versioning. You need a way to perform precise, multi-step revisions without the agent rewriting the entire file or losing track of your original intent.
What it does
- Routes tasks between chat and Canvas based on artifact complexity and revision needs.
- Executes targeted inline edits using precise coordinate prompts to prevent meaning drift.
- Performs structured code reviews and debug traces with line-by-line logical analysis.
- Handles prose restructuring and style polishing while strictly preserving factual claims.
- Converts code and documents across formats while maintaining original naming and error handling.
Frameworks & tools
ChatGPT Canvas (GPT-4o+), HTML/CSS/JS for live previews, and Markdown for document structuring.
Why this beats prompting it yourself
Standard prompts often result in the AI rewriting everything, which wastes tokens and introduces bugs. This skill uses a disciplined, multi-pass workflow that separates restructuring from polishing, ensuring no data is dropped during the transition. It enforces a "validate then move" protocol that prevents the hallucination of new claims or code behaviors.
Use cases
- Refactoring a complex Python script into TypeScript while keeping all docstrings intact.
- Restructuring a technical report into new sections without losing specific data points.
- Adding debug logging to every error path in a large codebase using a specific format.
- Performing a style polish on an essay to reduce word count by 20% without adding new arguments.
Known limitations
Does not support direct UI automation like clicking buttons. Limited to tasks within the ChatGPT environment; no external network requests or file system writes.
No direct UI control (clicking/dragging). No external API calls or file system access. Previews are limited to the Canvas sandbox environment. Paragraph drift can occur in very long documents.
How to use
- 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
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 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 1 month ago
Permissions
No special permissions declared or detected