compound-engineering-plugin-mirror/plugins/compound-engineering/skills/ce-commit-push-pr/references/pr-description-writing.md

PR Description Writing

How to resolve the right commit range and compose a PR title and body. Loaded on demand by ce-commit-push-pr callers — do not load earlier.

Step Pre-A resolves the commit range, diff, and (for existing PRs) the current PR body. Steps A through H assume Pre-A's outputs are in context.

---

Step Pre-A: Resolve the PR commit range and diff

Determine which commits and diff the description should cover. Run this first; Steps A and beyond assume the commit list and full diff are in context.

Mode

• Current-branch mode. Describe HEAD vs the repo's default base. Used when the caller has no explicit PR reference (Step 6 of ce-commit-push-pr's full workflow; description-only mode without a PR ref).
• PR mode. Describe a specific PR's commit range. Used by DU-3 (the existing PR on the current branch) and description-only mode (the user pasted a PR URL/number). The PR ref may be a bare number, #NN, pr:NN, or a full URL — the caller passes it to gh pr view <ref> directly. (gh pr view accepts numbers and URLs natively; #NN and pr:NN need the # or pr: stripped before passing.)

Resolve PR metadata (PR mode and current-branch-with-existing-PR)

gh pr view [<ref>] --json baseRefName,headRefOid,baseRefOid,headRepository,headRepositoryOwner,isCrossRepository,url,body,state

If the returned state is not OPEN, report "PR <number> is <state>; cannot generate description" and exit gracefully — do not invent a description.

Detect the base branch and remote

For PR mode (or current-branch mode with an existing PR), use baseRefName from the metadata above. For current-branch mode without an existing PR, resolve in priority order:

1. Caller-supplied base (if the caller passed base:<ref>) — use verbatim. The ref must resolve locally.
2. Repo default branch — git rev-parse --abbrev-ref origin/HEAD, strip origin/ prefix.
3. GitHub metadata — gh repo view --json defaultBranchRef --jq '.defaultBranchRef.name'.
4. Common names — try main, master, develop, trunk in order via git rev-parse --verify origin/<candidate>.

If none resolve, ask the user to specify the base branch.

For the base remote, parse owner/repo from the PR URL (or use the repo of the current checkout in current-branch mode without a PR) and match against git remote -v fetch URLs (handle both git@github.com:owner/repo and https://github.com/owner/repo forms; strip .git). The matching remote name is <base-remote>. If no remote matches (fork-based PR with no upstream remote), do NOT default to origin — origin would diff against the wrong base. Skip directly to Case B.

Case A — base remote matched, attempt local git

Only fetch when the refs aren't already local. Skipping the fetch when refs resolve locally preserves offline / restricted-network / expired-auth workability — a common case when re-running the skill on a branch that's already been fetched recently.

PR_HEAD_SHA=<headRefOid>   # for current-branch mode, use HEAD instead

if ! git rev-parse --verify <base-remote>/<baseRefName> >/dev/null 2>&1 \
  || ! git rev-parse --verify $PR_HEAD_SHA >/dev/null 2>&1; then
  git fetch --no-tags <base-remote> <baseRefName> $PR_HEAD_SHA
fi

MERGE_BASE=$(git merge-base <base-remote>/<baseRefName> $PR_HEAD_SHA) \
  && echo '=== COMMITS ===' && git log --oneline "$MERGE_BASE..$PR_HEAD_SHA" \
  && echo '=== DIFF ===' && git diff "$MERGE_BASE...$PR_HEAD_SHA"

For current-branch mode, $PR_HEAD_SHA is HEAD which is always local, so only the base ref needs fetching. Using the explicit $PR_HEAD_SHA in downstream commands avoids FETCH_HEAD's multi-ref ordering problem.

If git merge-base itself fails (shallow clone with insufficient history, or genuinely unrelated histories), do not press on — fall through to Case B and let the API path produce the diff and commits.

Case A inner fallback — SHA fetch rejected

Some GHES configurations disallow fetching non-tip SHAs. If the SHA fetch is rejected (PR mode only — current-branch mode uses HEAD which is always local), fall back to fetching the PR head via refs/pull/<number>/head:

git fetch --no-tags <base-remote> "refs/pull/<number>/head"
PR_HEAD_SHA=$(awk '/refs\/pull\/[0-9]+\/head/ {print $1; exit}' "$(git rev-parse --git-dir)/FETCH_HEAD")

Then re-run the merge-base + log + diff with the new $PR_HEAD_SHA.

Case B — API only

Use Case B when:

• No local remote matches the PR's base repo (fork-PR, cross-repo PR), OR
• Case A's fetches all fail (offline, restricted network, expired auth, GHES quirks blocking both SHA and refs/pull/N/head).

Skip local git entirely:

gh pr diff <ref>
gh pr view <ref> --json commits --jq '.commits[] | [.oid[0:7], .messageHeadline] | @tsv'

Note in the user-facing output that the API fallback was used.

Empty range

If the resulting commit list is empty (HEAD already merged into the base, or the branch has no unique commits), report "No commits to describe" and exit gracefully — do not invent a description.

---

Step A: Classify commits

Scan the commit list and classify each commit:

• Feature commits -- implement the PR's purpose (new functionality, intentional refactors, design changes). These drive the description.
• Fix-up commits -- iteration work (code review fixes, lint fixes, test fixes, rebase resolutions, style cleanups). Invisible to the reader.

When sizing the description, mentally subtract fix-up commits: a branch with 12 commits but 9 fix-ups is a 3-commit PR.

---

Step B: Decide on evidence

Decide whether to include an evidence section in the body.

Evidence is possible when the diff changes observable behavior demonstrable from the workspace: UI, CLI output, API behavior with runnable code, generated artifacts, or workflow output.

Evidence is not possible for:

• Docs-only, markdown-only, changelog-only, release metadata, CI/config-only, test-only, or pure internal refactors
• Behavior requiring unavailable credentials, paid/cloud services, bot tokens, deploy-only infrastructure, or hardware not provided

Decision logic:

1. Existing PR body contains a ## Demo or ## Screenshots section with image embeds: preserve it verbatim unless the user's focus asks to refresh or remove it. Include the preserved block in the body.
2. No existing evidence block: omit the evidence section entirely unless the caller already captured evidence and passed it in (e.g., a ce-demo-reel URL).

Do not label test output as "Demo" or "Screenshots". Place any preserved evidence block before the Compound Engineering badge.

---

Step C: Frame the narrative

Articulate the PR's narrative frame:

1. Before: What was broken, limited, or impossible? (One sentence.)
2. After: What's now possible or improved? (One sentence.)
3. Scope rationale (only if 2+ separable-looking concerns): Why do these ship together? (One sentence.)

This frame becomes the opening. For small+simple PRs, the "after" sentence alone may be the entire description.

---

Step D: Size the change

Assess size (files, diff volume) and complexity (design decisions, trade-offs, cross-cutting concerns) to select description depth:

Change profile	Description approach
Small + simple (typo, config, dep bump)	1-2 sentences, no headers. Under ~300 characters.
Small + non-trivial (bugfix, behavioral change)	Short narrative, ~3-5 sentences. No headers unless two distinct concerns.
Medium feature or refactor	Narrative frame (before/after/scope), then what changed and why. Call out design decisions.
Large or architecturally significant	Narrative frame + up to 3-5 design-decision callouts + 1-2 sentence test summary + key docs links. Target ~100 lines, cap ~150. For PRs with many mechanisms, use a Summary-level table to list them; do NOT create an H3 subsection per mechanism. Reviewers scrutinize decisions, not inventories — the diff and spec files carry the detail. If you find yourself writing 10+ subsections, consolidate to a table.
Performance improvement	Include before/after measurements if available. Markdown table works well.

When in doubt, shorter is better. Match description weight to change weight. Large PRs need MORE selectivity, not MORE content.

---

Step E: Apply writing principles

Writing voice

If the repo has documented style preferences in context, follow those. Otherwise:

• Active voice. No em dashes or -- substitutes; use periods, commas, colons, or parentheses.
• Vary sentence length. Never three similar-length sentences in a row.
• Do not make a claim and immediately explain it. Trust the reader.
• Plain English. Technical jargon fine; business jargon never.
• No filler: "it's worth noting", "importantly", "essentially", "in order to", "leverage", "utilize."
• Digits for numbers ("3 files"), not words ("three files").

Writing principles

• Lead with value: Open with what's now possible or fixed, not what was moved around. The subtler failure is leading with the mechanism ("Replace the hardcoded capture block with a tiered skill") instead of the outcome ("Evidence capture now works for CLI tools and libraries, not just web apps").
• No orphaned opening paragraphs: If the description uses ## headings anywhere, the opening must also be under a heading (e.g., ## Summary). For short descriptions with no sections, a bare paragraph is fine.
• Describe the net result, not the journey: The description covers the end state, not how you got there. No iteration history, debugging steps, intermediate failures, or bugs found and fixed during development. This applies equally when regenerating for an existing PR: rewrite from the current state, not as a log of what changed since the last version. Exception: process details critical to understand a design choice.
• When commits conflict, trust the final diff: The commit list is supporting context, not the source of truth. If commits describe intermediate steps later revised or reverted, describe the end state from the full branch diff.
• Explain the non-obvious: If the diff is self-explanatory, don't narrate it. Spend space on things the diff doesn't show: why this approach, what was rejected, what the reviewer should watch.
• Use structure when it earns its keep: Headers, bullets, and tables aid comprehension, not mandatory template sections.
• Markdown tables for data: Before/after comparisons, performance numbers, or option trade-offs communicate well as tables.
• No empty sections: If a section doesn't apply, omit it. No "N/A" or "None."
• Test plan — only when non-obvious: Include when testing requires edge cases the reviewer wouldn't think of, hard-to-verify behavior, or specific setup. Omit when "run the tests" is the only useful guidance. When the branch adds test files, name them with what they cover.
• No Commits section: GitHub already shows the commit list in its own tab. A Commits section in the PR body duplicates that without adding context. Omit unless the commits need annotations explaining their ordering or shipping rationale.
• No Review / process section: Do not include a section describing how the reviewer should review (checklists of things to look at, process bullets). Process doesn't help the reviewer evaluate code. Call out specific non-obvious things to scrutinize inline with the change that warrants it.

Visual communication

Include a visual aid only when the change is structurally complex enough that a reviewer would struggle to reconstruct the mental model from prose alone.

The core distinction — structure vs. parallel variation:

• Use a Mermaid diagram when the change has topology — components with directed relationships (calls, flows, dependencies, state transitions, data paths). Diagrams express "A talks to B, B talks to C, C does not talk back to A" in a way tables cannot.
• Use a markdown table when the change has parallel variation of a single shape — N things that share the same attributes but differ in their values. Tables express "option 1 costs X, option 2 costs Y, option 3 costs Z" cleanly.

Architecture changes are almost always topology (components + edges), so Mermaid is usually the right call — a table of "components that interact" loses the edges and becomes a flat list. Reserve tables for genuinely parallel data: before/after measurements, option trade-offs, flag matrices, config enumerations.

When to include (prefer Mermaid, not a table, for architecture/flow):

PR changes...	Visual aid
Architecture touching 3+ interacting components (the components have directed relationships — who calls whom, who owns what, which skill delegates to which)	Mermaid component or interaction diagram. Do not substitute a table — tables cannot show edges.
Multi-step workflow or data flow with non-obvious sequencing	Mermaid flow diagram
State machine with 3+ states and non-trivial transitions	Mermaid state diagram
Data model changes with 3+ related entities	Mermaid ERD
Before/after performance or behavioral measurements (same metric, different values)	Markdown table
Option or flag trade-offs (same attributes evaluated across variants)	Markdown table
Feature matrix / compatibility grid	Markdown table

When in doubt, ask: "Does the information have edges (A → B) or does it have rows (attribute × variant)?" Edges → Mermaid. Rows → table. Architecture has edges almost by definition.

When to skip any visual:

• Sizing routes to "1-2 sentences"
• Prose already communicates clearly
• The diagram would just restate the diff visually
• Mechanical changes (renames, dep bumps, config, formatting)

Format details:

• Mermaid (default for topology). 5-10 nodes typical, up to 15 for genuinely complex changes. Use TB direction. Source should be readable as fallback.
• ASCII diagrams for annotated flows needing rich in-box content. 80-column max.
• Markdown tables for parallel-variation data only.
• Place inline at point of relevance, not in a separate section.
• Prose is authoritative when it conflicts with a visual.

Verify generated diagrams against the change before including.

Numbering and references

Never prefix list items with # in PR descriptions — GitHub interprets #1, #2 as issue references and auto-links them.

When referencing actual GitHub issues or PRs, use org/repo#123 or the full URL. Never use bare #123 unless verified.

Applying user focus

If the user provided a focus hint (e.g., "emphasize the benchmarks", "this needs to read as a migration not a feature"), incorporate it alongside the diff-derived narrative. Treat focus as steering, not override: do not invent content the diff does not support, and do not suppress important content the diff demands simply because focus did not mention it. When focus and diff materially disagree (e.g., focus says "include benchmarking" but the diff has no benchmarks), surface the conflict to the user rather than fabricating content.

---

Step F: Compose the title

Title format: type: description or type(scope): description.

• Type is chosen by intent, not file extension or diff shape. feat for new functionality, fix for a bug fix, refactor for a behavior-preserving change, docs for doc-only, chore for tooling/maintenance, perf for performance, test for test-only. Where fix and feat could both seem to fit, default to fix: a change that remedies broken or missing behavior is fix even when implemented by adding code. Reserve feat for capabilities the user could not previously accomplish. The user may override.
• Scope (optional) is the narrowest useful label: a skill/agent name, CLI area, or shared area. Omit when no single label adds clarity.
• Description is imperative, lowercase, under 72 characters total. No trailing period.
• If the repo has commit-title conventions visible in recent commits, match them.

Breaking changes use ! (e.g., feat!: ...) or document in the body with a BREAKING CHANGE: footer. Do not apply either marker without explicit user confirmation — they trigger automated major-version bumps in some release tooling.

---

Step G: Compose the body

Assemble the body in this order:

1. Opening -- the narrative frame from Step C, at the depth chosen in Step D. Under a heading (e.g., ## Summary) if the description uses any ## headings elsewhere; a bare paragraph otherwise.
2. Body sections -- only the sections that earn their keep for this change: what changed and why, design decisions, tables for data, visual aids when complexity warrants. Skip empty sections entirely.
3. Test plan -- only when non-obvious per the writing principles. Omit otherwise.
4. Evidence block -- only the preserved or freshly captured block from Step B, if one exists. Do not fabricate or placeholder.
5. Compound Engineering badge -- append a badge footer separated by a --- rule. Skip if regenerating an existing body that already contains the badge.

Badge:

---

[![Compound Engineering](https://img.shields.io/badge/Built_with-Compound_Engineering-6366f1)](https://github.com/EveryInc/compound-engineering-plugin)
![HARNESS](https://img.shields.io/badge/MODEL_SLUG-COLOR?logo=LOGO&logoColor=white)

Harness lookup:

Harness	LOGO	COLOR
Claude Code	claude	D97757
Codex	(omit logo param)	000000
Gemini CLI	googlegemini	4285F4

Model slug: Replace spaces with underscores. Append context window and thinking level in parentheses if known. URL-encode literal parens as %28 and %29 — unencoded ( and ) inside a markdown image URL break some commit-message parsers (notably release-please's conventional-commits parser, which fails the whole commit on encountering them and silently skips it from the changelog). Examples: Opus_4.6_%281M,_Extended_Thinking%29, Sonnet_4.6_%28200K%29, Gemini_3.1_Pro.

---

Step H: Compression pass

Before applying, re-read the composed body and apply these cuts:

• If any body section restates content already in the ## Summary, remove it. The Summary plus the diff should carry the reader.
• If "Testing" or "Test plan" has more than 2 paragraphs, compress to bullets.
• If a "Commits" section enumerates the commit log, remove it — GitHub shows it in its own tab.
• If a "Review" or process-oriented section lists how to review, remove it. Move any truly non-obvious review hints inline with the relevant change.
• If the body has 5+ H3 subsections that each describe one mechanism, consolidate them into a single table row per mechanism under one header. Reserve prose H3 callouts for 2-3 genuine design decisions.
• If the body exceeds the sizing-table target by more than 30%, compress the longest non-Summary section by half.

Value-lead check. Re-read the first sentence of the Summary. If it describes what was moved around, renamed, or added ("This PR introduces three-tier autofix..."), rewrite to lead with what's now possible or what was broken and is now fixed ("Document reviews previously produced 14+ findings requiring user judgment; this PR cuts that to 4-6.").

Large PRs benefit from selectivity, not comprehensiveness.