add-task
by NeoLabHQadd-task creates draft task files in .specs/tasks/draft/ from a user request, preserving intent, type, and dependencies. This add-task skill is useful for Project Management workflows, spec-driven task tracking, and repo-aware task drafting before planning or implementation.
This skill scores 68/100, which means it is worth listing but with caveats. Directory users get a fairly triggerable task-creation workflow that preserves user intent and writes to a specific draft folder, but they should expect some rough edges in discoverability and install readiness because the repo shows placeholder markers, no install command, and no supporting reference files.
- Clear trigger and output target: creates a draft task file in .specs/tasks/draft/ from a task title or description.
- Operational workflow is spelled out with directory setup, input analysis, and task classification steps.
- Good agent leverage: it specifies filename, type classification, dependencies, and intent preservation, reducing guesswork versus a generic prompt.
- Repo signals some incompleteness: placeholder markers are present, which suggests the skill may still be partially unfinished.
- No install command or support files are provided, so setup and adoption may require extra manual interpretation.
Overview of add-task skill
What add-task does
The add-task skill creates a draft task file in .specs/tasks/draft/ from a user’s original request. It is useful when you want a consistent task artifact before planning or implementation, not just a freeform note. This makes the add-task skill a good fit for Project Management workflows where tasks need to be captured with a clear title, type, dependencies, and preserved intent.
Who should install it
Install add-task if your team uses spec-driven task tracking, separates drafting from execution, or wants agents to turn messy requests into structured task files. It is most valuable when task intake matters: product ops, engineering leads, AI-assisted PM workflows, and repo maintainers who need a repeatable draft-to-todo process.
Why it stands out
The main value of add-task is its structure. It does more than summarize a request: it helps classify the task, identify dependency files, and place the output in the right directory. That reduces guesswork when compared with a generic prompt that may forget naming conventions, task type, or folder conventions.
How to Use add-task skill
Install and activate add-task
Use the skill in the context of the NeoLabHQ/context-engineering-kit plugin set, then point the agent at the task request you want drafted. The install flow shown in the repository is:
npx skills add NeoLabHQ/context-engineering-kit --skill add-task
After installation, the agent should receive the user’s task title or description plus any dependency task files. The add-task install step matters because the skill expects repo-aware paths and folder setup, including .specs/tasks/draft/.
Write input the skill can parse
The best add-task usage is a compact but specific request. Include the outcome, scope, and any known dependencies. Strong input looks like:
Add validation to form inputsCreate login rate limiting; depends on auth cleanup taskRefactor task creation flow for mobile users
Weak input is vague and forces the model to guess:
Improve appFix stuffMake tasks better
If the request has implied type, say it. If not, let the skill infer it from the verbs and scope.
Read the right files first
Start with SKILL.md, because it defines the role, goal, input, and instructions. Then inspect any repo files that control task naming, workflow, or conventions. In this repository preview, SKILL.md is the primary source; there are no support folders like rules/ or references/, so the main risk is missing local conventions in your own workspace rather than missing documentation here.
Workflow that produces better drafts
Use add-task when the goal is to convert a rough request into a draft that can later move to todo or implementation. First, provide the exact task intent and dependencies. Next, let the skill create the draft file in the expected folder. Finally, review the output for title clarity, dependency accuracy, and whether the description preserved the user’s meaning without overediting it.
add-task skill FAQ
Is add-task for Project Management or coding?
Both, but the primary use is Project Management for engineering work. It creates a task artifact that can feed planning and implementation. If you only need a quick reminder note, a normal prompt is simpler.
Do I need to know the repository structure first?
No, but you should know whether your project uses .specs/tasks/ or a similar convention. If your repo has different task folders, adapt the output path before relying on the default add-task behavior.
When should I not use add-task?
Do not use it when you already have a fully specified implementation ticket, when the task should go straight to coding, or when your repo does not use draft task files at all. In those cases, a direct prompt or another workflow may be faster.
Is the skill beginner-friendly?
Yes, if you can describe a task in one sentence and understand basic dependency naming. The main learning curve is not the AI step; it is knowing what your team expects in the draft file.
How to Improve add-task skill
Give the skill fewer blanks to fill
The biggest quality gain comes from better source input. Include the action, target area, and any blockers. For example, Add password reset email retry handling; depends on email queue stability task is much better than Improve reset flow. This helps add-task preserve intent and assign dependencies correctly.
Check title, type, and dependencies
Most weak outputs come from three issues: titles that are too broad, type labels that are too generic, and missing dependency links. Review those first. If the title is not verb + specific object, rewrite it. If the type is uncertain, add a clarifying phrase before re-running the add-task skill.
Iterate by refining the prompt, not the file
If the first draft is off, improve the input instead of manually patching everything. Tell the agent what to preserve, what to reclassify, and what files it should depend on. This is the fastest add-task guide pattern for getting a draft that is ready for downstream planning.