gofenix/nex-agent/github-issue-executor
Claude Code and agent skill by gofenix/nex-agent
- community
Checking that the SKILL.md file is still in the repository.
Handle GitHub issue or PR comments that mention NexAgent, using the comment and repository context to decide whether to start, continue, retry, stop, report status, or ask for clarification.
Install
npx vskill@latest install gofenix/nex-agent/github-issue-executorSKILL.md
GitHub Issue Executor
You received a GitHub comment mention event. Your job is to understand the user's natural-language comment in the current issue or PR context, then take the appropriate GitHub and local execution steps.
This is not a chat request, Feishu task, Nex personal task request, or keyword command parser. Do not look for rigid commands such as run, revise, retry, or stop. Infer intent from the comment plus issue, PR, registry, branch, work_dir, and recent execution context.
Forbidden:
- Do not use Feishu or
lark-cli. - Do not use the internal
tasktool. - Do not use
executor_statusorexecutor_dispatch. - Do not implement issue changes yourself with read/edit/write tools.
- Do not inspect the repository at length before deciding whether coding work is needed.
- Do not execute risky or ambiguous requests; ask a concise GitHub comment question instead.
Step 1: Parse Metadata
Extract these fields from the incoming metadata when present:
- event_type
- issue_url, issue_number, issue_title, issue_body
- repo: GitHub repo such as
gofenix/nex-agent - work_dir: disposable local checkout path
- branch: branch to use inside the disposable checkout
- pr_url: existing pull request URL
- comment_body, comment_url, comment_author
- registry entry or previous run summary
If the visible message conflicts with metadata, trust metadata.
Step 2: Decide The Next Move
Use the comment and GitHub context to decide what to do.
Examples:
- The issue has no open PR and the user asks NexAgent to handle the issue: start implementation work.
- There is an open PR and the user says something is wrong or asks for a change: continue on the existing branch/PR.
- The last execution failed and the user says the blocking condition is fixed or asks to try again: retry the relevant step.
- The user asks to stop, pause, or abandon current work: stop local execution if possible and leave a clear GitHub comment.
- The user asks for progress: inspect local registry/PR state and report status.
- The user's intent is unclear or high-risk: ask one concise clarifying question in GitHub.
Do not require or invent an action metadata field.
Step 3: Prepare Work Directory
If coding work is needed:
- Validate
work_direxists or can be prepared from metadata. - Refuse to run if
work_diris under/Users/fenix/github/or/Users/fenix/repos/; those are long-lived checkouts. - Read the target repo's local instructions, especially
AGENTS.md, before dispatching. - Switch/create the branch from metadata, or continue the existing PR branch.
- Inspect the issue and PR enough to build the execution prompt:
Step 4: Dispatch Coding Work
Dispatch implementation work with the opencode_run tool. This is mandatory for code changes. Do not implement the issue yourself.
The nested opencode prompt must include:
- the issue title/body
- the existing PR context, if any
- the user's latest comment
- a clear instruction not to commit, push, or create a PR
- the narrowest useful verification expectation
Use:
Step 5: Publish Results
After opencode_run returns:
- Inspect the result JSON, stdout/stderr log, git diff summary, and local git diff.
- Run any missing narrow verification yourself.
- If changes are valid, commit and push the branch.
- Create a PR if none exists, or update the existing PR branch.
- Comment on the issue or PR with a concise result and verification summary.
- Record useful state in
tasks/github_items.json, includingrepo,issue_number,pr_url,work_dir,branch,last_comment_url, andlast_check.
If the user asked to stop or the intent is unclear, do not run opencode_run; leave a GitHub comment explaining the current state or asking for the missing decision.
Security
- Security scanPassed, 100 of 100Pattern checks on the skill's files.
- Model reviewNot runA model review is scheduled for the next rescan.
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingpriv/skills/github-issue-executor/SKILL.md
Popularity
- Repository stars
- 64
- Repository forks
- 4
- 7-day trend score
- 21
Stars and forks are for the whole GitHub repository, not this skill alone. Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/gofenix/nex-agent/github-issue-executor)