af-clone-issue

name	af-clone-issue
description	Copies a GitHub issue (title, body, and all comments in chronological order) from an upstream repository into a fork or other target repo using the GitHub CLI and optional jq.
argument-hint	https://github.com/{owner}/{repo}/issues/{number}

Copy GitHub issue to another repo

If .agent-fox/steering.md exists in the project root, read it and follow any directives found there before proceeding.

GitHub does not offer “duplicate this issue to my fork.” This workflow creates a new issue on the target repository and replays upstream comments as new comments with attribution.

Prerequisites

• gh installed and authenticated (gh auth login).
• Read access to the source issue; permission to open issues on the target repo.
• jq recommended for parsing comment JSON in the shell. If jq is missing, install it or expand the JSON manually (same fields: user.login, created_at, body).

Parse URLs

From an issue URL https://github.com/OWNER/REPO/issues/N:

• Source: SRC_OWNER, SRC_REPO, ISSUE_NUM (= N).

From a repo URL https://github.com/OWNER/REPO or shorthand OWNER/REPO:

• Target: DST_OWNER, DST_REPO.

Reject references that resolve to a pull request (issues and PRs share numbers). After fetching the issue JSON, if pull_request is present, stop and tell the user this number is a PR, not an issue.

Step 1 — Fetch the source issue

gh api "repos/${SRC_OWNER}/${SRC_REPO}/issues/${ISSUE_NUM}"

Confirm the response is an issue (pull_request must be absent). Note title, body (may be null), and html_url.

Step 2 — Fetch all comments (paginated)

gh api "repos/${SRC_OWNER}/${SRC_REPO}/issues/${ISSUE_NUM}/comments" --paginate

The CLI returns a single JSON array of all pages. Sort by created_at ascending before replaying (e.g. pipe through jq 'sort_by(.created_at)').

Step 3 — Build the new issue body

Prepend a provenance block, then the original body (use empty string if body was null):

> **Copied from** [SRC_OWNER/SRC_REPO#N](ISSUE_HTML_URL) on YYYY-MM-DD.

[original issue body]

Use UTC date for YYYY-MM-DD. Write the full body to a temp file (e.g. mktemp + heredoc or redirect) so multiline content is preserved.

Step 4 — Create the issue on the target repo

gh issue create \
  --repo "${DST_OWNER}/${DST_REPO}" \
  --title "TITLE_FROM_SOURCE" \
  --body-file /path/to/body.md

Capture stdout: it contains the new issue URL. Parse the new issue number from the path (.../issues/<num>).

Step 5 — Replay each comment

For each comment in chronological order, post one new comment on the new issue:

**@LOGIN** commented at `CREATED_AT_ISO`

COMMENT_BODY

• LOGIN from user.login (use unknown if missing).
• CREATED_AT_ISO from created_at.
• COMMENT_BODY from body (empty string if null).

Write each combined block to a temp file and run:

gh issue comment NEW_ISSUE_NUM \
  --repo "${DST_OWNER}/${DST_REPO}" \
  --body-file /path/to/comment.md

Shell pattern (gh + jq)

Set variables, then:

ISSUE_JSON=$(gh api "repos/${SRC_OWNER}/${SRC_REPO}/issues/${ISSUE_NUM}")
if echo "$ISSUE_JSON" | jq -e 'has("pull_request")' >/dev/null 2>&1; then
  echo "Ref is a pull request, not an issue." >&2
  exit 1
fi

TITLE=$(echo "$ISSUE_JSON" | jq -r '.title')
UPSTREAM_URL=$(echo "$ISSUE_JSON" | jq -r '.html_url')
RAW_BODY=$(echo "$ISSUE_JSON" | jq -r '.body // ""')
TODAY=$(date -u +%Y-%m-%d)

BODY_FILE=$(mktemp)
{
  printf '> **Copied from** [%s/%s#%s](%s) on %s.\n\n' \
    "$SRC_OWNER" "$SRC_REPO" "$ISSUE_NUM" "$UPSTREAM_URL" "$TODAY"
  printf '%s' "$RAW_BODY"
} >"$BODY_FILE"

OUT=$(gh issue create --repo "${DST_OWNER}/${DST_REPO}" --title "$TITLE" --body-file "$BODY_FILE")
rm -f "$BODY_FILE"
NEW_NUM=$(echo "$OUT" | sed -n 's/.*\/issues\/\([0-9][0-9]*\).*/\1/p')

COMMENTS_JSON=$(gh api "repos/${SRC_OWNER}/${SRC_REPO}/issues/${ISSUE_NUM}/comments" --paginate)
echo "$COMMENTS_JSON" | jq -c 'sort_by(.created_at) | .[]' | while read -r row; do
  login=$(echo "$row" | jq -r '.user.login // "unknown"')
  created=$(echo "$row" | jq -r '.created_at')
  cbody=$(echo "$row" | jq -r '.body // ""')
  CF=$(mktemp)
  printf '**@%s** commented at `%s`\n\n%s' "$login" "$created" "$cbody" >"$CF"
  gh issue comment "$NEW_NUM" --repo "${DST_OWNER}/${DST_REPO}" --body-file "$CF"
  rm -f "$CF"
done

echo "$OUT"

On macOS and Linux, date -u +%Y-%m-%d works for TODAY.

What this does not copy

Labels, milestones, assignees, projects, reactions, edits, or links to PRs. Label names often do not exist on the fork; recreate manually if needed.

Limits

GitHub enforces maximum sizes for issue and comment bodies; extremely large threads may need splitting (rare).

Optional dry run

Before gh issue create, show the user the resolved SRC_*, DST_*, title, body length, and comment count. Use gh api .../comments --paginate | jq 'length' for count after sort if needed.