ce-impactlisted
Install: claude install-skill relusion/vg-sdlc-claude-plugins
# Impact
**Invocation input:** Change to analyze: $ARGUMENTS
Given a **proposed change described in plain text** — typically the title +
description of an issue tracker work item (Azure DevOps, Jira, GitHub) pasted in —
report what that change would **touch in this codebase**: affected components, the
blast radius, an approximate sizing hint, similar prior work, and the questions the
description leaves unanswerable. Output is **grounded, cited findings** that a human
reads during refinement — never a go/no-go verdict.
This tool is **read-only**, **plan-free**, and **repo-agnostic**. It is **not** part
of the plan/spec/implement pipeline and consumes none of those artifacts. It is the
*forward-looking* sibling of `/ce-ask` (which answers questions about code as it is) and
reuses `/ce-ask`'s scoping discipline plus `/ce-plan`'s reachability / blast-radius
vocabulary — it does not re-derive that reasoning.
**The tool is tracker-agnostic by design.** It takes plain text and knows nothing
about Azure DevOps or any API. The most common source is an ADO work item, but the
skill never reads from or writes to a tracker — see *Cross-cutting rule — One-Way*.
## Runtime Inputs
- **Change description (required):** the text of the proposed change. Usually a
pasted work item (title + description, and acceptance criteria if present), but any
natural-language change request works. Examples:
- *"Add CSV export to the orders report"*
- *"Bug: login is slow when a user has many se