Orchestrate multi-phase repository implementation workflows with artifact files under .ai/<task-slug>.
npx skills add https://github.com/Bush2021/ayugram --skill task-think
Please help me install this Agent Skill. Skill: task-think Source: https://github.com/Bush2021/ayugram/tree/main/.codex/skills/task-think Command: npx skills add https://github.com/Bush2021/ayugram --skill task-think
This skill orchestrates a structured, multi-phase implementation workflow for a repository, producing artifact files under .ai/<task-slug> and optional fresh codex exec runs per phase.
Run the task-think skill to drive a complete phase-by-phase workflow for a repository, generating artifacts under .ai/<slug> and logging phase results.
High-intent search queries and answers about installing and using this skill.
You orchestrate a multi-phase implementation workflow by running a structured process that guides context gathering, planning, implementation, verification, and review in distinct phases. This approach enforces artifact checks and logs phase results for traceability.
The best way to manage workflow artifacts is to generate and maintain files like inputs.md, context.md, plan.md, implementation.md, and review.md automatically under the .ai/<task-slug> directory during each distinct workflow phase.
Yes, you can keep the main session context clean by supporting per-phase child codex exec runs. This isolates execution while still producing the required artifact files and logging the phase results back for traceability.
No external dependencies are required beyond your repository tooling. The structured process operates independently, enforcing artifact checks and logging phase results without needing additional external packages or frameworks.
You should use a phased implementation workflow when tasks require iterative improvements across context gathering, planning, implementation, verification, and review, ensuring traceability through generated artifact files without polluting the main session context.
Task-think generates artifact files under the .ai/<task-slug> directory to maintain a structured process and provide traceability. It logs phase results and checks artifacts to guide iterative improvements throughout the workflow.
Snowflake-LabsA concise summary of your roleplay character: name,</name> personality, and/or behavior, formatted as a single string sentence. Typically不超过50字，此处省略
mietheScore composed artifacts against a blueprint rubric and output scores, issues, and pass.
khanhbkqtDelegate scoped implementation, research and bug-fix tasks to Gemini or Codex worker agents via DEFINECOMPOSESPAWNREVIEWREPORT protocol scripts/templates/spawn-agent.sh with adjustable timeout, agent type, and approval modes.
DowwieCoordinate Task Decomposition Protocol workflows from specification to execution.
