[prompt-analysis] Copilot PR Prompt Analysis - 2026-10-21

[github-actions[bot]]

🤖 Copilot PR Prompt Pattern Analysis - 2026-10-21

Summary

Analysis Period: Last 30 days
Total PRs: 633 | Merged: 526 (83.1%) | Closed: 101 (16.0%) | Open: 6 (0.9%)

Merge rate among completed PRs: 83.9%. No PR had an empty body.

Prompt Categories and Success Rates (title keyword matching)

Category	Total	Merged	Success Rate
Bug Fix	76	66	86.8%
Feature Addition	98	79	80.6%
Refactoring	16	13	81.3%
Documentation	33	26	78.8%
Testing	17	14	82.4%
Other/uncategorized	393	328	83.5% (of completed)

Prompt Analysis

✅ Successful Prompt Patterns

• Average prompt length: 175.7 words
• Prompts referencing specific files (.go, .cjs, .md, .yml...): 128 merged / 18 closed = 87.7% vs 82.7% (398/83) for those without
• Action verbs: fix, align, record, make, add

Examples:

1. PR Align agentic-workflows guidance link format #70862: Align agentic-workflows guidance link format → Merged
2. PR Fix runner-label indentation in custom jobs #70840: Fix runner-label indentation in custom jobs → Merged
3. PR Make admission jobs opt-in through frontmatter or aw.json #70763: Make admission jobs opt-in through frontmatter or aw.json → Merged

❌ Unsuccessful Prompt Patterns

• Average prompt length: 198.2 words (open PRs avg 108)
• Closed PRs skew toward broad, design-heavy features

Examples:

1. PR Support team-configurable default models #70349: Support team-configurable default models → Closed
2. PR Generate GitHub-native workflow lockfiles during compilation #70345: Generate GitHub-native workflow lockfiles during compilation → Closed
3. PR Use artifact-backed credit accounting for daily code metrics #69988: Use artifact-backed credit accounting for daily code metrics → Closed

Key Insights

• Specific scope wins: narrow fixes (86.8%) outperform broad features (80.6%); file references add ~5 points.
• Longer ≠ better: closed prompts are ~13% longer than merged ones, suggesting sprawling scope hurts.
• Docs/feature variance is modest; differences are small-sample and should be read cautiously.

Recommendations

1. DO: Reference concrete files/functions and the expected outcome.
2. DO: Keep one focused change per prompt; split large features.
3. AVOID: Broad architectural asks without acceptance criteria.

Historical Trends

Date	PRs	Merge Rate
2026-10-21	633	83.9%
2026-10-20	622	83.5%
2026-10-19	611	83.6%
2026-10-18	602	83.6%
2026-10-17	604	83.8%

Trend: Merge rate is stable around 83–84%.

---

Generated by Copilot PR Prompt Analysis (Run: 39672821523)

Generated by 🔍 Copilot PR Prompt Pattern Analysis · copilot · auto · 14.3 AIC · ⌖ 0.717 AIC · ⊞ 11.6K · ◷

• expires on Oct 22, 2026, 12:49 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 #71221.