github-issue
Skill1 starsChanged 25 days agoMITView on GitHub
Generates well-structured GitHub issue markdown for feature/task issues (title, Summary, Motivation, Acceptance Criteria, optional Out of Scope) and bug reports (title, Description, Steps to reproduce, Expected behavior, Environment, Additional context). Triggers on /github-issue, "write a github issue", "draft an issue", "generate issue markdown", "create a github issue for", "write a bug report", "file a bug".
20 other projects use a version of this.
Adopted by
Most recently taken up by
- StephanMeijer/skillstheir own version, first seen Nov 6, 2026
- stempeck/agentfactorytheir own version, first seen Oct 30, 2026
- enkr1/claude-code-skillstheir own version, first seen Oct 20, 2026
- razbakov/skillstheir own version, first seen Oct 17, 2026
- rabble/nostrvinetheir own version, first seen Oct 9, 2026
- divinevideo/divine-mobiletheir own version, first seen Oct 8, 2026
Used with
Projects with it also use zeroclaw, github-pr and 2 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
openagenSkillzeroclaw
Help users operate and interact with their ZeroClaw agent instance — through both the CLI (`zeroclaw` commands) and the REST/WebSocket gateway API. Use this skill whenever the user wants to: send messages to ZeroClaw, manage memory or cron jobs, check system status, configure channels or providers, hit the gateway API, troubleshoot their ZeroClaw setup, build from source, or do anything involving the `zeroclaw` binary or its HTTP endpoints. Trigger this even if the user just says things like 'check my agent status', 'schedule a reminder', 'store this in memory', 'list my cron jobs', 'send a message to my bot', 'set up Telegram', 'build zeroclaw', or 'my bot is broken' — these are all ZeroClaw operations.
Also in 3 projects that use this
ncklrsSkillgithub-pr
Also in 6 projects that use this
SentrySkillskill-creator
Create new agent skills following the Agent Skills specification. Use when asked to "create a skill", "add a new skill", "write a skill", "make a skill", "build a skill", or scaffold a new skill with SKILL.md. Guides through requirements, writing, registration, and verification.
Also in 4 projects that use this
SentrySkillcommit
Create commit messages following Sentry conventions. Use when committing code changes, writing commit messages, or formatting git history. Follows conventional commits with Sentry-specific issue references.
Also in 3 projects that use this
Other versions20 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 20
- zeroclaw-labs/zeroclaw★ 33ksame name
- openagen/zeroclaw★ 1.9ksame name
- rhesis-ai/rhesis★ 404same name
- divinevideo/divine-mobile★ 265same name
- rabble/nostrvine★ 265same name
- theonlyhennygod/zeroclaw★ 103same name
- giulio333/ClaudeLens★ 67same name
- devstefancho/skills★ 13same name
- yantrikos/yantrikclaw★ 9same name
- Cogni-AI-OU/cogni-ai-agent-skills★ 4same name
10 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository12 other files it gives its agents.
- a11y-auditSkill
- backend-conventionsSkill
- decision-deckSkill
- frontend-conventionsSkill
- k8s-conventionsSkill
- new-nebari-appSkill
- new-projectSkill
- pr-reviewSkill
- start-devSkill
- talk-deckSkill
- tool-evalSkill
- web-app-auditSkill
The file
We split this skill into the rules it gives an agent
11 rules. 0 are common, 8 are distinctive, and 3 are very specific to this project.
0
Common
Most of the 19 edited versions we read keep these.
8
Distinctive
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 22
If $ARGUMENTS is empty, ask the user one short question: "What should the issue be about?" Then proceed once they answer.
- Distinctive: line 26
If a flag was given, use it.
- Distinctive: line 28
Bug: when the request describes something that is broken, crashes, errors, regresses, or behaves differently from what was expected — look …
- Distinctive: line 31
If it is genuinely ambiguous (e.g.
- Very specific: line 35
Read the two skill-relative files for the chosen type (resolve paths relative to ~/.claude/skills/github-issue/):
- Distinctive: line 46
Only if it makes the issue materially better and the info is cheap to get:
- Distinctive: line 50
Do not block on questions.
- Distinctive: line 56
Write the issue using exactly these sections, in this order:
- Very specific: line 66
Write the issue using exactly these sections, in this order (mirrors the bug_report.yml form fields, rendered as plain markdown):
- Distinctive: line 77
Use the section headings verbatim for the chosen template.
- Distinctive: line 79
Keep bullets tight — a single claim each.
“Common” is measured: kept by at least 6 in 10 of the 19 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
You are generating markdown for a GitHub issue. Produce clean, well-structured markdown that follows the selected template exactly. Your job is to turn the user's description into a complete, reviewable issue — not to ask a long series of questions first.
Input
$ARGUMENTS is a free-form description of the feature, task, change, or bug the issue should cover. It may be a single sentence ("add Biome for linting", "deploy crashes when config has no cluster block") or a detailed paragraph. The user may also reference the current repository, paste a stack trace, or paste existing notes.
An optional leading flag forces the issue type:
--bug— use the bug report template.--feature(or--task) — use the feature/task template.
If $ARGUMENTS is empty, ask the user one short question: "What should the issue be about?" Then proceed once they answer.
Step 1 — Pick the issue type
If a flag was given, use it. Otherwise infer the type from the description:
- Bug when the request describes something that is broken, crashes, errors, regresses, or behaves differently from what was expected — look for words like "bug", "broken", "fails", "crash", "error", "traceback", "regression", "doesn't work", "wrong", "unexpected", or a pasted stack trace / error message.
- Feature/task (the default) for everything else — new functionality, refactors, tooling changes, chores, docs, improvements.
If it is genuinely ambiguous (e.g. "the login page is slow" could be a bug or a performance improvement), default to feature/task and state the choice in the assumptions line after the output. Do not ask.
Step 2 — Read the template and example
Read the two skill-relative files for the chosen type (resolve paths relative to ~/.claude/skills/github-issue/):
| Type | Template | Example |
| Feature/task | assets/issue-template.md | assets/example.md |
| Bug | assets/bug-template.md | assets/bug-example.md |
The template gives the exact section structure; the example shows the expected tone, density, and formatting.
Step 3 — Gather just-enough context (optional)
Only if it makes the issue materially better and the info is cheap to get:
- If the working directory is a git repo and the issue is about this codebase, you may inspect relevant files (e.g.
package.json, config files, existing tooling, the code path named in a traceback) to make the content concrete and accurate. Keep this lightweight — a few reads, not a full audit. - For bugs, if the user pasted a traceback or error, use it verbatim in the Description; if they named a command that failed, put that exact command in Steps to reproduce. You may fill in the Environment from cheap local facts (e.g.
uname,python --version, the project's version file) only when the bug is clearly about the current machine/repo — otherwise leave placeholders for the user to fill in. - Do not block on questions. Make reasonable, clearly-stated assumptions instead. If a critical detail is genuinely ambiguous, ask at most one focused question.
Step 4 — Generate the markdown
Feature/task issues
Write the issue using exactly these sections, in this order:
## <Title>— H2 heading. Action-oriented and concise (e.g. "Add Biome for Formatting & Linting"). This is the issue title; everything below is the body.### Summary— 1–3 sentences. What the issue proposes and its scope. Link tools/libraries with markdown links. State which parts of the codebase it touches.### Motivation— bullet list. One specific reason or benefit per bullet.### Acceptance Criteria— GitHub task-list checkboxes (- [ ]). Usually 3–5 items. Each one states an outcome that defines "done" and can be checked, in a short line. Describe what must be true, not how to build it: leave out implementation steps, file-by-file changes, config keys, and edge-case enumerations unless one is essential to the requirement. The person picking up the issue decides the details.### Out of Scope— optional; omit it by default. Include it only when there is a genuine exclusion a reader would otherwise reasonably assume is included (e.g. an adjacent language, platform, or follow-up the user explicitly deferred). One or two bullets. Never invent exclusions just to fill the section.
Bug reports
Write the issue using exactly these sections, in this order (mirrors the bug_report.yml form fields, rendered as plain markdown):
## [BUG] - <Title>— H2 heading with the literal[BUG] -prefix. The title names the defect concretely: the command/page, the symptom, and the trigger where known (e.g. "nebi deployfails withKeyError: 'namespace'when--configomits theclusterblock"). Avoid vague titles like "Deploy broken".### Description— 1–3 sentences: what happened, where, and the impact. If there is an error message or traceback, include it verbatim in a fenced code block directly under the prose.### Steps to reproduce— numbered list. Each step is a single concrete action (exact command, page, click). The last step states the observed result.### Expected behavior— what should have happened instead, specific enough that a fix can be verified against it.### Environment— bullet list of- Key: valuelines. Always includeOSand the project's version (name it after the project, e.g.Nebi version); add others that plausibly matter (browser, runtime version, install method, cluster tooling). Use placeholders like<fill in>rather than guessing values you cannot verify.### Additional context— bullet list: logs, screenshots, related issues/PRs, workarounds tried, narrowing observations. Omit the section entirely if there is nothing useful to add.
Formatting rules (both types)
- Use the section headings verbatim for the chosen template.
- Wrap commands, filenames, config keys, and package names in backticks.
- Keep bullets tight — a single claim each. No filler.
- Fill every required section. Out of Scope (features) and Additional context (bugs) are optional — omit them when there is nothing real to say.
Step 5 — Output
By default, print the generated markdown directly in your reply inside a single fenced ```markdown code block so the user can copy it cleanly. Note the chosen issue type (if inferred) and any assumptions you made in one short line after the block.
Then offer the follow-up actions (do not perform them unless the user asks):
- Save to a file (e.g.
gh-issue.md). - Create the issue with
gh:gh issue create --title "<title>" --body-file <file>(the title is the##heading text without the##; the body is everything below it). For bug reports, also pass--label "type: bug"if that label exists in the target repo. Only run this if the user explicitly asks andghis available.
Important Notes
- Read the template and example first — do not generate from memory; match their structure and tone.
- Pick one template and stick to it — never mix feature sections (Motivation, Acceptance Criteria) into a bug report or bug sections (Steps to reproduce, Environment) into a feature issue.
- Title comes from the
##heading — when creating the issue viagh, strip the leading##and use the rest as--title(keep the[BUG] -prefix for bugs); use the remaining sections as the body. - Don't invent reproduction details — for bugs, only list steps and environment facts the user gave you or you verified locally; use placeholders for the rest.
- Don't over-interrogate — generate a strong draft from what you're given, state assumptions, and let the user refine.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
Reviews can't be read right now.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

