Execute planned tasks one at a time with Codex or Gemini, verifying and committing each.
npx skills add https://github.com/buvis/home --skill work-buvis
Please help me install this Agent Skill. Skill: work Source: https://github.com/buvis/home/tree/main/.claude/skills/work Command: npx skills add https://github.com/buvis/home --skill work-buvis
Implementing planned work can be error-prone and slow; this Skill provides a disciplined, one-task-at-a-time workflow that uses Codex or Gemini to execute tasks and commits after each completion.
Initiate the one-task-at-a-time workflow by selecting the first pending task, dispatching the appropriate agent, and committing after completing each task.
High-intent search queries and answers about installing and using this skill.
To execute planned tasks one at a time, this Skill dispatches a single task to Codex or Gemini, verifies the results with tests, and commits the changes before moving to the next task. This structured workflow ensures deterministic and auditable task management.
Automating task execution and version control commits is best handled by a trigger-driven workflow where an AI agent completes a task, runs tests for verification, and automatically commits the work. This approach minimizes errors and maintains a disciplined development pace.
Yes, you can use Codex for backend work and Gemini for frontend work within the same workflow. The Skill orchestrates end-to-end task execution by assigning each task to the appropriate AI agent and applying per-task verification before committing.
You start a trigger-driven workflow for implementing planned work by using specific prompts like 'work on tasks', 'implement tasks', 'start working', 'execute the plan', or 'do the work'. These triggers initiate the one-task-at-a-time execution loop.
Yes, this one-task-at-a-time workflow includes automated testing and verification. After an AI agent completes a dispatched task, the workflow applies explicit per-task verification steps to ensure the results meet requirements before committing the changes.
The limitation of using AI agents for structured task execution is that work must be planned and broken down into discrete tasks beforehand, as the workflow only processes one task at a time and requires explicit per-task verification before committing any changes.
sunLeeeExecute task files sequentially with human approval at each step.
deonmenezesCheck source line execution status against gcov .gcda data.
chiruu12Automate software deployment with Git branch management, testing, and pull requests.
etrExecute remaining tasks in isolated agent team sessions.
