github-issue-reply

A workflow for investigating and replying to GitHub issues, which are reports or requests attached to a software project. It checks the issue details and the latest published release before suggesting a maintainer response.

In plain words
What is it for?
Use it to review bug reports, feature requests, questions, and follow-ups; compare them with the latest release; write a restrained reply; choose suitable labels; and close issues that no longer need tracking.
Why use it?
It reduces the risk of giving an unsupported answer about whether a bug is fixed. It also helps keep issue labels and closure decisions consistent with the repository's existing practices.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/sundegan/agent-skills/github-issue-reply
Any agent
npx skills add sundegan/agent-skills --skill github-issue-reply
Clone the repo
git clone --depth 1 https://github.com/sundegan/agent-skills

Made for: Claude Code, Codex.

Per session 86 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,832 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00086 $0.01832
Opus 5 $0.00043 $0.00916
Sonnet 5 $0.00017 $0.00366
Haiku 4.5 $0.00009 $0.00183

Measured 2d ago against content hash ee4de874dd54, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

github-issue-reply scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

general-skills/github-issue-reply/SKILL.md · 98 lines

How it starts

The opening of the file, as written. The whole thing — 98 lines — stays where its author put it; the contents beside it link to each section on GitHub.

GitHub Issue Reply

Goal

Produce a concise maintainer reply that sounds human, acknowledges useful feedback without exaggerated praise, and accurately tells the reporter whether the latest published version addresses the issue. Apply appropriate labels and close conclusively resolved issues that no longer need tracking.

Workflow

  1. Establish the repository and issue:
    • Read the complete issue title, body, labels, comments, linked pull requests, and referenced issues.
    • Identify the reported behavior, expected behavior, requested feature, affected version, platform, and reproduction details.
    • Determine the repository from the current Git remote or the issue URL. Do not guess when multiple repositories are possible.
    • Inspect the repository's existing labels and recent issue-labeling conventions before choosing labels.
  2. Inspect the latest published release:
    • Prefer gh release view --json tagName,name,publishedAt,url,isDraft,isPrerelease,body and gh release list when GitHub CLI access is available.
    • Exclude drafts. Treat prereleases as the latest only when the issue concerns prerelease users or the user explicitly asks for them.
    • Inspect release notes, changelogs, merged pull requests, commits, and source code as needed. Search using the issue's feature names, error text, affected components, and likely synonyms.
    • Use the repository's actual release URL. Prefer the specific release page because it contains version details and assets; use the repository's /releases/latest URL only when a stable redirect is more appropriate.
  3. Determine the result from evidence:
    • Mark implemented or fixed only when release notes, a merged change included in that release, or the released source clearly establishes the connection.
    • Mark possibly related when evidence is suggestive but incomplete. Ask the reporter to test without stating that the issue is fixed.
    • Mark not yet implemented when the latest release does not contain the requested behavior.
    • If there is no published release, say so plainly and do not invent a version or download link.
  4. Draft the reply in the repository's established language and maintainer style:
    • Determine the reply language from the issue title, body, and full comment thread, not from the repository language alone.
    • Reply in Chinese when the reporter uses Chinese and the discussion is entirely or effectively Chinese.
    • Reply in English when the reporter and discussion use English.
    • Prefer English when participants use a mixture of Chinese and English, even if the original report is Chinese, so the reply remains accessible to everyone in the thread.
    • Follow an explicit language request from the user when one is provided.
    • Preserve code, error messages, product names, release names, and established technical terms in their original form when translating them would reduce clarity.
    • Build the reply from the specific issue context rather than selecting or filling a reusable response template.
    • Refer to the concrete behavior, requested capability, platform, reproduction detail, or discussion point that matters to the resolution.
    • Thank or compliment the reporter only when the context supports it. Make any acknowledgment specific to what was useful, such as a reproduction, log, diagnosis, or design suggestion.
    • Decide the order and contents from the situation. Include only the relevant combination of resolution, explanation, version, release link, upgrade detail, verification request, workaround, remaining limitation, and closure notice.
    • State the verified status clearly, but do not force every reply into the same opening, paragraph structure, or closing sentence.
    • Invite further testing or feedback only when it is useful. Tailor the requested follow-up to the unresolved risk instead of asking everyone to “try the latest version and let us know.”
  5. Choose labels:
    • Reuse the repository's existing labels and naming conventions. Do not create a new label merely because a generic label name seems appropriate.
    • Apply the smallest useful set, such as the repository's equivalents of bug, enhancement, question, duplicate, fixed, released, needs-info, or a relevant component/platform label.
    • Preserve existing labels unless they are clearly incorrect or obsolete. Do not remove labels unrelated to the current resolution.
    • Do not apply mutually contradictory status labels.
  6. Decide whether to close:
    • Close the issue only when the reported bug or requested feature is conclusively fixed or implemented in a published release and no further maintainer tracking is needed.
    • Keep the issue open when evidence is incomplete, the reporter still needs to verify an uncertain fix, the change exists only on an unreleased branch, additional work remains, or the issue tracks broader follow-up work.
    • A request to try a conclusively verified released fix does not by itself require keeping the issue open. Tell the reporter they can comment or open a new issue if the problem persists, following repository convention.
    • Use an appropriate close reason when the platform supports it and repository convention is clear. Do not classify a completed fix as “not planned.”
  7. Execute the issue update:
    • When the user asks to reply to or handle an issue, post the comment, apply the selected labels, and then close it if the closure criteria are met.
    • Prefer gh issue comment, gh issue edit --add-label, and gh issue close when GitHub CLI access is available.
    • Perform the comment before closure so the resolution is visible in the issue timeline.
    • Re-read the resulting issue state to verify the comment, labels, and open/closed status.
    • Return only a draft and make no GitHub changes when the user explicitly asks for a draft, template, example, or review.
    • If authentication or permissions prevent an operation, report exactly which comment, label, or closure action was not completed.
  8. Report the result:
    • Include the issue URL, posted reply status, labels added or retained, and whether the issue was closed.
    • Return only the publishable reply when the user asks for a draft.
    • When investigation details are useful, put a short evidence note after the draft, clearly separated from the publishable text.

Read the full file on GitHub · 98 lines

Files

What ships with it

1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 98 lines · 86 tokens per session scan A ee4de874dd54

Subscribe to this mod's changes

github-issue-reply is a skill published in the GitHub repository sundegan/agent-skills (2 stars, last pushed 1mo ago), licensed MIT. It adds 86 tokens to every session and 1,832 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other skills, from other repositories

systematic-debugging

Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.

obra/superpowers · 21 tokens

brainstorming

You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.

obra/superpowers · 37 tokens

auto-perf-optimize

Run agent-driven VS Code performance or memory investigations. Use when asked to launch Code OSS, automate a VS Code scenario, run the Chat memory smoke runner, capture renderer heap snapshots, take workflow screenshots, compare run summaries, or drive a repeatable scenario before heap-snapshot analysis.

microsoft/vscode · 62 tokens

chat-perf

Run chat perf benchmarks and memory leak checks against the local dev build or any published VS Code version. Use when investigating chat rendering regressions, validating perf-sensitive changes to chat UI, or checking for memory leaks in the chat response pipeline.

microsoft/vscode · 51 tokens

chat-pet-sprite-creation

Use when creating or changing VS Code chat pet sprite art, sprite sheets, state animations, eye treatments, Stable/Insiders variants, or pet transitions under src/vs/workbench/contrib/chat/browser/widget/media/chatPet.

microsoft/vscode · 53 tokens

cpu-profile-analysis

Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…

microsoft/vscode · 71 tokens