Create GitHub Issues from Unstructured Data

This skill transforms messy input, such as error logs, voice notes, and screenshots, into clean, actionable GitHub issues with reproduction steps, impact, and evidence. It writes issues as markdown files in an /issues/ directory at the repo root, named YYYY-MM-DD-short-description.md.

Its output template has eight sections: Summary (a one-line description), Environment (product/service, region or version, browser/OS if relevant), Reproduction Steps (numbered), Expected Behavior, Actual Behavior, Error Details (in a code block when applicable), Visual Evidence (a reference to attached screenshots or GIFs), Impact (a severity rating with a brief explanation), and Additional Context.

Its guidelines: be crisp, with no fluff and every word adding value; extract structure from chaos, since voice dictation and raw notes often bury facts in casual language; infer missing context, using conversation or memory to fill in specifics when a user mentions something like "same project" or "the dashboard"; placeholder sensitive data with tags like [PROJECT_NAME] or [USER_ID]; and match severity to impact using four tiers - Critical (service down, data loss, security issue), High (a major feature broken with no workaround), Medium (a feature impaired but a workaround exists), and Low (a minor inconvenience or cosmetic issue). Image or GIF attachments are referenced inline.

Two worked examples show the transformation: a rambling voice dictation about an agent that "just failed silently" becomes a structured issue with Summary, Environment, numbered Reproduction Steps, Expected/Actual Behavior, a High impact rating, and the note that three retries were needed; a pasted "Error: PERMISSION_DENIED... Code: 403" message becomes a structured issue with an Error Details code block and a High impact rating noting it is a regression from a previously working state.

Use it when you have unstructured bug input, such as pasted errors, support notes, screenshots, or voice dictation, and need to turn it into a clean GitHub issue with a summary, reproduction steps, expected versus actual behavior, impact, and attachment references.

Inputs are error logs, voice notes, screenshots, or rough bug descriptions in any casual form. Outputs are a markdown file in /issues/ named by date and short description, structured into the eight-section template with severity-rated impact and sensitive data placeholdered.

It is for developers and support engineers who receive bug reports in unstructured form (voice, chat, screenshots, pasted errors) and need them turned into consistent, developer-ready GitHub issues without manually rewriting each one.