report

name	report
description	File a machine self-diagnostic report as a tagged GitHub issue. When a tool, daemon, hook, skill, or agent hangs, errors, returns garbage, or behaves wrong, open a `machine-report`-labeled issue on the machine plugin repo so every plugin user's reports funnel to the maintainer. Requires `gh` logged in. Write-side only — it logs, it does not fix. Trigger: "/report", "report a problem", "this tool is broken", "file a machine report", "log this hang/error/hiccup".
metadata	{"version":"2.0.0","category":"workflow","status":"active","updated":"2026-06-17","tags":"report, bug, hang, broken, diagnostic, self-fix, tooling, hiccup, github, issue"}

/report — file a machine defect as a tagged GitHub issue

The machine should learn from its own friction. When a tool hangs, a daemon won't come up, a hook misfires, a skill does the wrong thing, or any part of the machine hiccups — don't just work around it silently. File a report.

Reports are GitHub issues, not local files. A local file under /.machine/ is only fixable by someone who forked the machine and runs your exact setup — useless to the maintainer and to every other install. An issue on the machine plugin repo funnels every user's reports to one place, where they can actually be triaged and fixed for everyone.

This skill is write-only. It opens the issue. It does not triage or fix — that is /resolve, which only runs in the machine repo itself.

Requirements

gh must be installed and authenticated. Check gh auth status. If it fails, do not fall back to a local file — tell the operator to run gh auth login, report that the issue could not be filed, and continue the original task. No login, no report.

When to file

File a report when the machine itself misbehaves:

• a tool/MCP call hangs, times out, or returns an error that isn't your input's fault
• a daemon (kern, hub) won't launch, drops, or returns nonsense
• a hook fires wrong, blocks, or never fires
• a skill or agent does something clearly off-spec
• a documented capability doesn't work as written
• repeated friction that a machine change would remove

Do not use it for: ordinary code bugs in the target project (those are normal work), or feature ideas (those are brainstorm/drill). Issues are for the machine's own defects.

Where reports go

Always the machine plugin repo, never the user's current repo — pass --repo explicitly so the issue lands upstream regardless of the user's cwd remote:

yesitsfebreeze/machine

Labels (the "tags"): machine-report always, plus severity:<severity> (blocker | major | minor | annoyance). Ensure both labels exist first — create them on demand so the flow works on a fork too.

Issue format

Title: one factual line naming the problem.

Body: keep it short — what broke, not prose.

- **severity:** blocker | major | minor | annoyance
- **area:** <tool / daemon / hook / skill / agent name>

## What happened
<2-4 lines: what you did, what the machine did instead.>

## Expected
<1-2 lines: what should have happened.>

## Evidence
<exact command/tool call + the error/output. A few lines, not a dump.>

## Context
<repo, OS, branch, agent role, anything that helps reproduce.>

## Suspected cause / fix (optional)
<only if you actually have a lead — otherwise omit.>

How to run

1. Check auth. gh auth status. If it fails, stop (see Requirements).

2. Dedupe. Search open reports for the same problem: gh issue list --repo yesitsfebreeze/machine --label machine-report --state open --search "<keywords>". If a matching issue exists, comment on it (gh issue comment <url> --body ...) instead of opening a near-duplicate — single source of truth applies to issues too.

3. Pick an area and a severity. Fill Evidence from the actual tool output you saw — quote it, don't paraphrase from memory.

4. Ensure labels, then open the issue (substitute <title>, <severity>, and write the body to a temp file or pass it inline):

   REPO=yesitsfebreeze/machine
   gh label create machine-report --color B60205 \
     --description "Filed by /report — machine self-diagnostic" --repo "$REPO" 2>/dev/null || true
   gh label create "severity:<severity>" --color CCCCCC --repo "$REPO" 2>/dev/null || true
   gh issue create --repo "$REPO" \
     --title "<title>" \
     --body "<body in the format above>" \
     --label machine-report --label "severity:<severity>"
   

5. Confirm the issue URL (gh issue create prints it) back to the operator in one line. Then continue the original task — filing a report does not abort your work, you note it and move on.

The consume side (machine repo only)

/report only ever writes. To act on the backlog, run /resolve in the machine repo — it lists the open machine-report issues, fixes the top one at its root cause, and closes that issue. /resolve is repo-local to the machine and is not shipped to other installs.