Run a personal assistant that handles repo issues and groceries
I switched to oh-my-py and never looked back. Task: The author uses oh-my-py to orchestrate AI agents for coding and personal tasks. For coding, one agent opens a repo issue, while another writes code, runs tests, and updates the master branch. For personal tasks, the system executes grocery purchases using Bitwarden secrets and a limited virtual card. The workflow leverages parallel agents for non-overlapping tasks and uses specific models (GLM, DeepSeek) via OpenRouter for planning and execution.; tools: custom-built or unnamed agent. Human role: The original post does not say how a person stayed involved.; limitations: The post does not detail specific operational constraints, costs, or failure modes. Reliance on third-party providers (OpenRouter) and specific model versions (GLM 5.3, DeepSeek Flash 4.1) introduces dependency risks. Security implications of granting agents access to payment methods and code repositories are not addressed. The claim of 'hundreds of hours' saved is anecdotal and unverified..
Tools
custom-built or unnamed agent
Published on Agent Work Atlas:
Overview
I switched to oh-my-py and never looked back.
Workflow
The author uses oh-my-py to orchestrate AI agents for coding and personal tasks. For coding, one agent opens a repo issue, while another writes code, runs tests, and updates the master branch. For personal tasks, the system executes grocery purchases using Bitwarden secrets and a limited virtual card. The workflow leverages parallel agents for non-overlapping tasks and uses specific models (GLM, DeepSeek) via OpenRouter for planning and execution.
Reported result
The source does not provide measured outcomes, efficiency metrics, or error rates. The author claims the ability to 'move hundreds of hours of work with one hour of input' but offers no data to substantiate this. No specific failures or successes are documented beyond the general description of the workflow.
Human role
The original post does not say how a person stayed involved.
Limitations
The post does not detail specific operational constraints, costs, or failure modes. Reliance on third-party providers (OpenRouter) and specific model versions (GLM 5.3, DeepSeek Flash 4.1) introduces dependency risks. Security implications of granting agents access to payment methods and code repositories are not addressed. The claim of 'hundreds of hours' saved is anecdotal and unverified.
Try it
Act as an automation engineer setting up a personal assistant workflow using oh-my-py. 1. **Context**: I want to automate two tasks: (A) Creating a GitHub issue and a corresponding PR for a bug fix in [replaceable_repo_url], and (B) Ordering groceries from [replaceable_grocery_service] with a budget of [replaceable_budget]. 2. **Tools**: Use oh-my-py with OpenRouter. Configure GLM 5.3 for planning and DeepSeek Flash 4.1 for execution. Use Bitwarden CLI for secrets. 3. **Steps**: - First, perform a dry run for the GitHub task. Generate the issue text and code changes but DO NOT push. Show me the diff and issue description for approval. - Once approved, execute the GitHub workflow: open issue, create branch, commit changes, run tests, and open PR. - For groceries, simulate the order process. List items and total cost. DO NOT charge the card. Wait for my explicit confirmation before attempting any real transaction. 4. **Output**: Provide a summary of actions taken, files changed, and any errors encountered. 5. **Safeguards**: If any step fails or requires permissions not granted, stop and ask for clarification. Never execute financial transactions without a prior 'dry run' and explicit user 'yes'.
Source reports are not independently verified. Publication date is the catalog publication date, not the original article date.