Skip to content
Skills Library
Management & product · Writing

issue

/issue

Use when filing or editing a GitHub issue, or saving a session's plan, analysis or visual onto one, in any repo.

Repository
loadout ↗
Licence
MIT
Last repository change
13.11.2026

Assessment and security

100/ 100Class: good
Format25 / 25
Frontmatter with name and description, valid name, description up to 1,024 characters, name matching the folder.
Discoverability25 / 25
Description explains when to use the skill, has a reasonable length and uses third person.
Content35 / 35
At least 150 words and at most 500 lines, headings, step lists, examples or code.
Hygiene15 / 15
No placeholders such as TODO, no Windows paths, with a licence.
Riskno signals detected in the scan
Files in this package
2
Scripts
0
HTML comments
0
URLs in SKILL.md
0
Deprecation

No signals detected. Repository changed on 13.11.2026, without references to retired models or APIs.

Licence

MIT, open: use and modify while retaining author attribution.

What it does and when to use it

Use when filing or editing a GitHub issue, or saving a session's plan, analysis or visual onto one, in any repo.

Original author description

Use when filing or editing a GitHub issue, or saving a session's plan, analysis or visual onto one, in any repo.

How it works

  1. Rules
  2. Artifacts
Original SKILL.md headings · 75 lines, 672 words
SKILL.md preview
---
name: issue
description: "Use when filing or editing a GitHub issue, or saving a session's plan, analysis or visual onto one, in any repo."
---

An issue is read by whoever picks the work up later, often an agent with none of your context.
Put what they must know on one screen, in this order: what is wrong or wanted, the proof, when it is done.
Everything else goes under `<details>`, and the session's durable work (a plan, a root-cause analysis, a visual) goes on the issue as **artifacts**, so it outlives the session.

````markdown
<One or two sentences: what is wrong or wanted, in the reader's words.>

## Evidence

- <the symptom as a run: the command and its output, `path:line`, or what the screen shows>
- **Suspected cause:** <one line, or "unknown">

## Done when

- <a check that passes: a command, a test, an observable result>

## Needs you

<Optional: one plain-English sentence per decision only the owner can make. No such item, no section.>

<details>
<summary>Context, reasoning, notes</summary>

<repro steps, what was tried, options weighed, related issues and PRs, what is not known>

</details>
````

## Rules

- **One screen.** At most about 15 lines outside `<details>`; one line per bullet, no paragraphs there.
- **Title is the ask,** imperative and specific: `Fail agent-eval on a refused key`, not `agent-eval bug`.
- **Evidence is runs, not adjectives,** as in the `pr` skill; a feature request's evidence is where its absence hurts.
- **Done when is checkable:** someone can tell done from not-done without asking you.
- **Needs you** holds only a real decision; never a no-op line.
- **Details hold the rest:** repro steps, logs beyond the key line, options and why, links.
  Several `<details>` blocks are fine when the topics differ; leave a blank line after `<summary>…</summary>` or GitHub shows raw Markdown.
- Search open issues first; add to a match (a comment in this same shape) instead of filing a twin.
- A repo's own issue template or labels win; they only rename these sections.
- Editing an existing issue follows the same shape; rewrite a wall of text into it rather than appending.

## Artifacts

The session's files vanish with it.
Before you leave an issue, keep the few that the next person would otherwise redo, as artifacts on it.

**Keep** (at most about three, each standing alone):

| What | `type` |
|---|---|
| An implementation plan | `plan` |
| A root-cause analysis, research notes, a decision write-up | `generic` (Markdown) |
| A self-contained HTML visual: diagram, report, prototype | `embedded_html` |
| A published page, a prototype, a thread elsewhere | `link` |

**Leave behind:** scratch files, full logs (quote the key line under Evidence), anything the repo or a rerun regenerates, and anything secret or private (a public repo's artifacts are public).

- **Name each for what it holds:** `Plan - fail fast on a refused key`, `RCA - retries ignored in eval` (names allow only letters, digits, spaces, `.`, `-`, `_`; a `:` or `/` is refused with 422).
- **Concise:** an artifact earns its place by saving the reader work; cut a session's narrative down to findings and decisions.
- **HTML stands alone:** inline CSS, images as `data:` URIs, no scripts or external requests (GitHub previews it in a sandbox and warns the reader about network calls).
- **Update, don't duplicate:** a newer version of the same plan is a `PATCH` to its artifact (GitHub keeps the history).
- The body points to an artifact only when it relies on it: `Plan: see the artifact "Plan - …"`.

```bash
# gh ≥ 2.103 has `gh issue artifact`; the REST calls work with any gh.
A=repos/<owner>/<repo>/issues/<n>/artifacts
gh api -X POST $A -f type=plan -f name="Plan - <what>" -F body=@plan.md   # -F reads the file; -f would post the path
gh api -X PATCH $A/<artifact_number> -F body=@plan.md
gh api $A --jq '.[] | "\(.number) \(.type) \(.name)"'                    # read back what landed
gh api $A/<artifact_number> --jq .artifact.body                          # one artifact reads as {artifact, versions}
```

Artifacts are in public preview: when the endpoint refuses, put the artifact's Markdown in a comment under `<details>` and say so.
Files in this package (2)
  • SKILL.md 4 KB
  • agents/openai.yaml 1 KB

Versions

RepositoryScoreLicence
This page · Paradiddle131/loadout100MIT
elalfner/relevo100MIT
ShieldBattery/ShieldBattery100MIT
chanmuzi/git-claw100MIT
adelrodriguez/skills100MIT
cyberyon-official/skillhead100MIT
bengous/claude-code-plugins100MIT
egregore-labs/egregore100MIT
andrewBezerra/concordia100MIT

Database snapshot: 13.11.2026 · package version bf02b602dade (SHA-256 digest of the entire package)

Downloads in the last 30 days: 0 · ranking

Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.

Similar skills

  • 90plan
    Step 1 of the SDLC. Turn a GitHub "Idea" issue into an intent file (intent/<issue#>-<slug>.md) via a PR, or sync an existing intent file when its issue has changed. Use when given an issue number to plan, or asked to check whether an issue and its intent file have drifted.
  • 90filing-issues
    Use when filing one or more GitHub issues for planned work — "file an issue for this", "turn this plan into issues", "open a tracking issue". Covers filing everything on skyf0xx/hedgehog regardless of which repo's code changes, whether to split a plan into a tracking issue plus…
  • 84plan
    Creates an implementation plan for a GitHub issue and posts it as a comment. Triggered when the user provides a GitHub issue number for planning.
  • 100issueplan
    Plan a task or GitHub issue on the current session model, then stop. A Plan subagent writes the plan by default; when the user asks for it, the main agent plans in the session with no subagent. With an issue, it posts the plan as an issue comment; with a task description and no…
  • 91issue-create
    Creates a GitHub issue from the project's issue form templates. Use when the user asks to open an issue, not to fix a bug.
  • 100git-issue-manage
    Transfer, pin, lock, or open a dev branch from a GitHub issue. Use when moving an issue between repos, pinning or locking it, or bulk-editing custom fields.

Feedback and reports

Your message goes to Bartosz Sokoliński. An email address is needed only for a reply. See the privacy policy.