create-item_skill

This skill creates MCP work items anchored to appropriate containers from conversation context, inferring type, priority, and notes for tracking.
  • Kotlin

166

GitHub Stars

1

Bundled Files

2 months ago

Catalog Refreshed

3 months ago

First Indexed

Readme & install

Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.

Installation

Preview and clipboard use veilstrat where the catalogue uses aiagentskills.

npx veilstrat add skill jpicklyk/task-orchestrator --skill create-item

  • SKILL.md7.0 KB

Overview

This skill creates MCP work items from conversation context with minimal friction. It infers title, type, priority, and scope, anchors the item into the appropriate container (Bugs, Features, Tech Debt, Observations, etc.), and pre-fills required notes when possible. Use it whenever a conversation surfaces a bug, feature idea, tech debt item, or any trackable observation.

How this skill works

The skill first infers intent from the conversation or provided arguments: title, type, priority, and whether a work tree is needed. It scans existing containers via an overview query to detect hierarchical, flat, or empty project structures and decides where to anchor the item. It reads available note schemas from the agent config to apply tags when a matching schema exists, creates the item (single or work tree) with inferred fields, and upserts any required queue notes using guidance from expectedNotes.

When to use it

  • When someone says “track this”, “log this bug”, or “create a task for” during conversation
  • When a bug, crash, or unexpected behavior is described and should be tracked persistently
  • When a new feature idea or enhancement is proposed that needs planning or subtasks
  • When tech debt, refactor work, or cleanup opportunities are identified
  • When an observation or workflow friction should be recorded for later action

Best practices

  • Supply concise context or an explicit title if available to avoid clarification questions
  • If multiple distinct subtasks are named, confirm to create a work tree; otherwise default to a single item
  • Keep required note content in the conversation or provide guidance so auto-filled notes are meaningful
  • Rely on the config.yaml note schemas to enforce gating; add a default schema if you want fallback behavior
  • If no project root exists, respond to the prompt about creating one to keep containers organized

Example use cases

  • A user reports a reproducible crash during file upload — create a Bug under Bugs and pre-fill reproduction steps
  • A designer proposes a new settings panel with three clear subtasks — create a Features work tree with those subtasks
  • A developer flags duplicate technical complexity to be cleaned up — create a Tech Debt item and tag accordingly
  • An agent notices a missing API capability during a run — create an Observation under Observations with captured evidence
  • A meeting note says “follow up on quota limits” — create an Action Item under Action Items with the follow-up instruction

FAQ

If no config file exists, the skill skips tagging and creates schema-free items. It will not apply tags that don’t match any schema key.

How does the skill decide between a single item and a work tree?

It creates a work tree only when the conversation explicitly names two or more distinct subtasks. Otherwise it defaults to a single item and asks a concrete yes/no if the scope is unclear.

What if the inferred container is wrong?

You can move the item after creation using an update operation to set the correct parentId. The skill will ask clarifying questions if type inference is uncertain.

Built by
VeilStrat
AI signals for GTM teams
© 2026 VeilStrat. All rights reserved.All systems operational
create-item skill by jpicklyk/task-orchestrator | VeilStrat