TinyFish Feedback

Collect the user's feedback, structure it, preview it, then file it. Nothing leaves the machine without the user seeing the exact text first.

Collect

Ask (briefly) for: what they were trying to do, what happened instead, and what they expected. For anything setup-, auth-, or connectivity-shaped, run npx -y @tiny-fish/cli@latest doctor and offer to attach its stdout JSON verbatim — it is schema-versioned and already redaction-safe (undeclared fields stripped on parse, messages authored rather than raw). Do not add fields, summarise it, or paste config contents alongside it. On exit 2 doctor produced no JSON — say so instead of attaching an empty report.

Structure

### What I was doing
…
### What happened
…
### Expected
…
### Environment
harness + version, CLI version (if known)
### Doctor report (optional)
`tinyfish doctor` stdout, verbatim

Preview gate

Show the complete issue body to the user and ask for an explicit yes before filing. Any edit they request happens before filing.

File

• Preferred: write the body and the title to files, then gh issue create --repo tinyfish-io/tinyfish-cookbook --title "$(cat "$title_file")" --body-file "$body_file" (only if gh is installed and authenticated). Both values come from the user's free-form text (the body also carries doctor's JSON), so neither may reach the shell as literal text — including in a title=… assignment, where backticks or $(…) are evaluated just the same. cat's output is not re-parsed. Never build the command as a string or run it through eval.
• Fallback: open a prefilled issue URL (https://github.com/tinyfish-io/tinyfish-cookbook/issues/new?title=…&body=…). Percent-encode both values — an unencoded # truncates the body and & splits it into junk parameters. URL length limits truncate long bodies anyway — if the body was truncated, tell the user and show the full text so they can paste the remainder.

This repo is public — remind the user of that in the preview if the report contains anything they typed free-form.