How do I install the Cv Handle Issue skill?

Run npx skillmds@latest add curvineio/cv-handle-issue in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Cv Handle Issue safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Cv Handle Issue?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Cv Handle Issue free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Cv Handle Issue?

curvineio (@curvineio) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Cv Handle Issue by curvineio — Agent Skill

Name: Cv Handle Issue
Author: curvineio

Cv Handle Issue

Fix a Curvine GitHub issue from analysis through plan approval to implementation, routing large features to cv-tasks-breakdown. Use when user asks to fix, resolve, or implement a specific issue URL or issue number.

curvineio 84e4216 3.0 KB Updated Sep 17, 2026

File contents

cv-handle-issue

End-to-end issue resolution with explicit user approval before coding.

When to Use

User provides an issue URL or issue number and asks to fix / implement / resolve it

Prerequisites

Valid issue reference (URL or #<number>)
gh auth status OK

Step 1: Read Issue

gh issue view <NUMBER> --json number,title,body,labels,state,url

Extract:

Bug vs feature (labels, title prefix)
Goal / Not Goal / Test Plan sections (if present)
Acceptance criteria and reproduction steps

If issue is closed or scope is unclear, stop and ask user.

Step 2: Analyze

Summarize for the user:

Item	Content
Type	bug / feature
Problem	What is wrong or what is requested
Affected modules	From issue body + codebase exploration
Risks	Breaking changes, concurrency, compatibility
Test strategy	Existing tests + new cases needed

Use deepwiki / codegraph as needed for CurvineIO/curvine.

Step 3: Propose Plan (No Coding Yet)

Present a plan:

## Proposed Approach

1. ...
2. ...

## Files to touch (estimate)
- `path/...` — ...

## Test plan
- ...

## Estimated size
~<N> lines / <M> files

Size routing

Estimate	Next step
≤ ~500 lines	Continue with this skill after user approval
> ~500 lines or multi-module feature	Use cv-tasks-breakdown; create sub-issues

Do not write code until the user explicitly approves the plan.

Iterate on the plan until approval. For ambiguous requirements, ask clarifying questions.

Step 4: Branch

After approval:

git fetch origin
git checkout main && git pull origin main
git checkout -b fix/<NUMBER>-<short-slug>   # or feat/...

Step 5: Implement

Minimal scope aligned with approved plan
Follow .cursor/rules/std-common/ai-workflow.mdc
Add/update tests per plan
Run make format and relevant tests before commit

If using task breakdown, complete one task / one commit at a time.

Step 6: Handoff to PR

When implementation and tests pass, use cv-create-pr:

PR title: <type>(<module>): <description> (#<NUMBER>)
Fill all PR body sections including Test verified

Rules

No coding without explicit user approval on the plan
Code change target: ≤ ~500 lines per issue unless broken down via cv-tasks-breakdown
Large features → sub-issues + task breakdown, not one monolithic PR

Break down large work → cv-tasks-breakdown
File new issue → cv-create-issue
Submit PR → cv-create-pr
Review feedback → cv-address-pr-review

---name: cv-handle-issuedescription: Fix a Curvine GitHub issue from analysis through plan approval to implementation, routing large features to cv-tasks-breakdown. Use when user asks to fix, resolve, or implement a specific issue URL or issue number.---# cv-handle-issueEnd-to-end issue resolution with explicit user approval before coding.## When to Use- User provides an **issue URL or issue number** and asks to fix / implement / resolve it## Prerequisites- Valid issue reference (URL or `#<number>`)- `gh auth status` OK## Step 1: Read Issue```bashgh issue view <NUMBER> --json number,title,body,labels,state,url```Extract:- Bug vs feature (labels, title prefix)- Goal / Not Goal / Test Plan sections (if present)- Acceptance criteria and reproduction stepsIf issue is **closed** or scope is unclear, stop and ask user.## Step 2: AnalyzeSummarize for the user:| Item | Content || ---- | ------- || Type | bug / feature || Problem | What is wrong or what is requested || Affected modules | From issue body + codebase exploration || Risks | Breaking changes, concurrency, compatibility || Test strategy | Existing tests + new cases needed |Use deepwiki / codegraph as needed for CurvineIO/curvine.## Step 3: Propose Plan (No Coding Yet)Present a plan:```markdown## Proposed Approach1. ...2. ...## Files to touch (estimate)- `path/...` — ...## Test plan- ...## Estimated size~<N> lines / <M> files```### Size routing| Estimate | Next step || -------- | --------- || ≤ ~500 lines | Continue with this skill after user approval || > ~500 lines or multi-module feature | Use [cv-tasks-breakdown](../cv-tasks-breakdown/SKILL.md); create sub-issues |**Do not write code until the user explicitly approves the plan.**Iterate on the plan until approval. For ambiguous requirements, ask clarifying questions.## Step 4: BranchAfter approval:```bashgit fetch origingit checkout main && git pull origin maingit checkout -b fix/<NUMBER>-<short-slug> # or feat/...```## Step 5: Implement- Minimal scope aligned with approved plan- Follow `.cursor/rules/std-common/ai-workflow.mdc`- Add/update tests per plan- Run `make format` and relevant tests before commitIf using task breakdown, complete **one task / one commit** at a time.## Step 6: Handoff to PRWhen implementation and tests pass, use [cv-create-pr](../cv-create-pr/SKILL.md):- PR title: `<type>(<module>): <description> (#<NUMBER>)`- Fill all PR body sections including **Test verified**## Rules- **No coding without explicit user approval** on the plan- Code change target: **≤ ~500 lines** per issue unless broken down via `cv-tasks-breakdown`- Large features → sub-issues + task breakdown, not one monolithic PR## Related- Break down large work → [cv-tasks-breakdown](../cv-tasks-breakdown/SKILL.md)- File new issue → [cv-create-issue](../cv-create-issue/SKILL.md)- Submit PR → [cv-create-pr](../cv-create-pr/SKILL.md)- Review feedback → [cv-address-pr-review](../cv-address-pr-review/SKILL.md)

Frequently asked questions