[prompt-analysis] Copilot PR Prompt Analysis - 2026-06-13

[github-actions[bot]]

🤖 Copilot PR Prompt Pattern Analysis — 2026-06-13

Summary

Analysis Period: Last 30 days | Total PRs: 1,000 | Merged: 792 (80%) | Closed: 204 (20%) | Open: 4

The overall success rate holds at 80%, consistent with the 79–81% range observed over the past week. Testing-focused prompts lead all categories at 88%, while WIP/checklist anti-patterns remain the most actionable levers for improvement.

Prompt Categories and Success Rates

Category	Total	Merged	Success Rate
🧪 Testing	52	46	88%
📝 Documentation	49	41	84%
�� Refactor	61	49	80%
✨ Feature	208	166	80%
🐛 Bug Fix	592	459	78%

Prompt Analysis

✅ Successful Prompt Patterns

Merged PR characteristics (avg 190 words):

• 85% include code blocks (vs 74% in closed PRs)
• 59% reference specific files

Top action verbs by success rate:

Verb	Total	Success Rate
remove	39	97%
add	139	83%
update	12	83%
doc	28	82%
fix	129	80%

Example merged prompts:

1. PR fix(unbloat-docs): fetch LFS objects during checkout to fix docs build #39145: docs/slides/github-agentic-workflows.pdf is stored in Git LFS. Without lfs: true, checkout leaves the LFS pointer in place, causing npm run build to fail... → Merged
2. PR Emit Copilot auth migration tip for copilot-requests workflows #39129: This change adds compiler guidance for Copilot-engine workflows to prefer GitHub Actions token-based inference... → Merged

❌ Unsuccessful Prompt Patterns

Closed PR characteristics (avg 216 words — 14% longer than merged):

Pattern	Total	Success Rate
WIP / Draft label	66	42%
Uncompleted checklist - [ ]	25	12%

Example closed prompts:

1. PR [WIP] Refactor to replace map[string]bool with map[string]struct{} #39009: [WIP] Refactor to replace map[string]bool with map[string]struct{} — [ ] Write Go AST rewriter tool... → Closed
2. PR [WIP] Fix Daily Agent of the Day Blog Writer workflow failure due to AI credits #38156: [WIP] Fix Daily Agent of the Day Blog Writer workflow failure due to AI credits... → Closed

Key Insights

• WIP = low confidence: WIP-tagged PRs succeed only 42% of the time vs 82% baseline. Don't submit until scope is fully defined.
• Uncompleted checklists predict failure: - [ ] items in PR bodies correlate with a 12% merge rate — nearly always indicates incomplete work.
• Code blocks matter: Merged PRs have code blocks 85% vs 74% of the time. Include diffs, snippets, or error output for better results.

Recommendations

1. DO include concrete code blocks, file paths, and error messages in prompts.
2. DO use unambiguous action verbs (remove, add, update) — remove achieves a 97% merge rate.
3. DO prefer testing or documentation prompts — both exceed 84% success.
4. AVOID WIP labels and incomplete checklists — 42% and 12% success rates respectively.
5. AVOID verbose bodies — keep under 200 words for best results.

Historical Trends

Date	Merged	Success Rate
2026-06-13	792	80%
2026-06-12	786	79%
2026-06-10	793	80%
2026-06-09	787	79%
2026-06-08	794	80%
2026-06-07	791	79%
2026-06-06	788	79%

Trend: Stable at 79–80%. No regression or improvement observed.

References: §27478909056

Generated by 🔍 Copilot PR Prompt Pattern Analysis · 475 AIC · ⌖ 28.8 AIC · ⊞ 20K · ◷

• expires on Jun 14, 2026, 1:12 PM UTC-08:00

[github-actions[bot]]

This discussion has been marked as outdated by Copilot PR Prompt Pattern Analysis.

A newer discussion is available at Discussion #39286.