Read task references, explore the codebase, and transition to plan mode.
npx skills add https://github.com/igor1309/skills --skill intake-igor1309
Please help me install this Agent Skill. Skill: intake Source: https://github.com/igor1309/skills/tree/main/intake Command: npx skills add https://github.com/igor1309/skills --skill intake-igor1309
Task intake for implementation and bug-fix assignments. Activates when the user says "/intake", "here's your task", "implement this", "fix this bug", "there's a bug", or provides a task reference (step number, PR, doc path) with instruction to understand before coding. The agent reads all referenced documents and explores the codebase, then demonstrates understanding in a structured playback before auto-transitioning to plan mode. For bug fixes, the playback includes a proposed failing test that proves the broken behavior. Prevents guessing, spec rewriting, and premature implementation.
Provide a task reference or explicit task description to initiate the intake process.
High-intent search queries and answers about installing and using this skill.
Task intake ensures an agent understands a bug report or feature request before coding by reading referenced documents, exploring the codebase, and demonstrating comprehension via structured playback before transitioning to plan mode.
The best way to align on a task reference before implementation planning is to initiate an intake process that reads all referenced documents, explores the codebase, and demonstrates understanding via playback to prevent guessing and spec rewriting.
You start the task intake process by providing a task reference like a PR, doc path, or explicit command, which triggers the agent to read references, explore the codebase, and demonstrate understanding before planning.
Yes, you can use intake for both feature requests and PR references, as it activates on explicit task intake commands to capture intent, explore the codebase, and demonstrate understanding via playback before planning.
Yes, the intake process generates a proposed failing test for bug fixes during the playback phase to prove the broken behavior before auto-transitioning to plan mode for the actual implementation.
An agent guesses or rewrites specs during premature implementation because it lacks grounded context, which task intake prevents by enforcing codebase exploration and structured playback before any planning begins.
davidabeyerExtract user-selected proposals into a structured brief for the /plan command.
alangebCapture and organize feature ideas into standardized markdown templates.
mathiasbourgoinGenerate neutral, intent-free research questions from task descriptions.
cmaennerDetect hardcoded credentials and secrets in source code and commit histories.
