Install
$ agentstack add skill-a1f-agent-templates-pr-breakdown ✓ scanned · ✓ verified — works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
About
PR Breakdown
Break the slice plan in a PRD issue into small PRs (~100–200 LOC, excluding tests), recorded as sub-rows in the same issue.
/pr-breakdown [--issue=N]
1. Find the issue and slice plan
2. Split each slice into PRs
3. Show it, wait for approval (gate)
4. Append to the SAME issue
Phase 1 — Find the issue and slice plan
Use --issue=N if given. Otherwise scan the conversation for the last PRD published: ... (issue #N) line; if there's none, ask. If you auto-detected it, confirm the number with the user. Read the body. If it has no ## Build plan section, stop and tell the user to run /to-issues first — don't invent a breakdown. If the Build plan says No slicing — single PR, emit one ### Slice 1 — (1 PR) row (derive `` from the PRD's What we're building) and skip to Phase 3.
Rules for splitting
- 100–200 LOC per PR, excluding tests. Over ~200 → split further; a tiny slice may be one PR.
- One cohesive, reviewable, demoable change per PR.
- Foundations first: reusable primitives before the PRs that wire them up.
- Note where PRs in a slice run in parallel.
- Rows only. Do NOT write coder prompts (module boundary, acceptance criteria, etc.) — those are generated later, one PR at a time.
Phase 2 — Draft sub-rows
Derive each ` from the slice's What`.
## PR breakdown
### Slice — ( PRs)
| PR | What | LOC | Done when |
|----|------|-----|-----------|
| n.1 | ... | ~150 | |
| n.2 | ... | ~120 | ... |
*n.1 ∥ n.2 (independent)*
End with Estimated total: ~N PRs.
Phase 3 — Approve (gate)
Show the full breakdown in chat. Wait for approval or edits before touching the issue.
Phase 4 — Append to the same issue
Keep canonical order PRD → ## Build plan → ## PR breakdown: replace the PR breakdown in place, preserving everything above it. The rebuild keys on the exact ## PR breakdown heading; tr -d '\r' guards CRLF; abort on an empty read:
ISSUE=; NEW=$(mktemp)
# ...write the new "## PR breakdown" section to "$NEW"...
CUR=$(gh issue view "$ISSUE" --json body -q .body | tr -d '\r')
[ -n "$CUR" ] || { echo "abort: empty issue body" >&2; exit 1; }
HEAD=$(printf '%s\n' "$CUR" | awk '/^## PR breakdown/{exit} {print}')
OUT=$(mktemp)
{ printf '%s\n\n' "$HEAD"; cat "$NEW"; } > "$OUT"
gh issue edit "$ISSUE" --body-file "$OUT"
rm -f "$NEW" "$OUT"
echo "PR breakdown appended to issue #$ISSUE"
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: a1f
- Source: a1f/agent-templates
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet — be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.