# Scope Guard

> Keep task execution within agreed scope. Use when tasks risk expanding into unrelated work, speculative features, or unbounded refactors. Track in-scope and out-of-scope items and protect delivery focus.

- Skill: `alubiama/scope-guard` (Agent Skill)
- Install (CLI): `npx skillmds@latest add alubiama/scope-guard`
- Raw SKILL.md: https://api.skillmd.com/api/skills/alubiama/scope-guard/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Productivity
- Author: Alubiama (https://skillmd.com/u/alubiama)
- Updated: 2026-09-22
- Page: https://skillmd.com/skills/alubiama/scope-guard

---


# Scope Guard

## Goal

Prevent scope creep and preserve delivery speed.

## Scope Workflow

1. Define `in_scope`, `out_of_scope`, and `deferred`.
2. Detect new requests that exceed current scope.
3. Propose options: include now, defer, or split into follow-up.
4. Continue execution only with approved scope.

## Trigger Conditions

Run this guard when:

- New feature requests appear mid-task.
- Time/cost constraints tighten.
- Refactor ideas exceed user intent.

## Output Contract

Always output:

1. `current_scope`
2. `scope_changes`
3. `decision` (`include`, `defer`, `split`)
4. `next_action`

## Guardrails

- Do not silently expand task scope.
- Keep backlog ideas in deferred list, not in active execution.