[prompt-analysis] Copilot PR Prompt Analysis - 2026-10-09

[github-actions[bot]]

🤖 Copilot PR Prompt Pattern Analysis - 2026-10-09

Summary

Analysis Period: Last 30 days
Total PRs: 575 | Merged: 482 (83.8%) | Closed: 87 (15.1%) | Open: 6 (1.0%)
Merge rate among completed PRs: 84.7%. No PRs had empty bodies.

Prompt Categories and Success Rates (by title keywords)

Category	Total (completed)	Merged	Success Rate
Bug Fix	90	78	86.7%
Feature Addition	82	64	78.0%
Refactoring	14	12	85.7%
Documentation	21	17	81.0%
Testing	11	9	81.8%
Other/uncategorized	351	302	86.0%

Prompt Analysis

✅ Successful Prompt Patterns

• Average body length: 160 words (merged)
• Bug-fix and refactor-style tasks merge most often; bodies referencing specific files (.go/.cjs/.md/.yml) merged 93/109 (85.3%) vs 389/460 (84.6%) without, so little difference.

Examples (merged):

1. Dispatch engine conformance through a priority-aware work queue #67222: Dispatch engine conformance through a priority-aware work queue → Merged
2. Bound AgentRx optimizer input to avoid prompt exhaustion #67133: Bound AgentRx optimizer input to avoid prompt exhaustion → Merged
3. Fix SEC-005 repository authorization conformance for Claim adapters #67132: Fix SEC-005 repository authorization conformance for Claim adapters → Merged

❌ Unsuccessful Prompt Patterns

• Average body length: 184 words (closed)
• Closed PRs skew toward feature work and cross-cutting/environment-specific concerns.

Examples (closed):

1. Disable uv Actions caching and isolate sandbox runtime paths #66958: Disable uv Actions caching and isolate sandbox runtime paths → Closed
2. Avoid GHES masking of cross-job workflow outputs #66678: Avoid GHES masking of cross-job workflow outputs → Closed
3. Prevent unresolved AI credit estimates from blocking workflows #66559: Prevent unresolved AI credit estimates from blocking workflows → Closed

Key Insights

• Feature additions have the lowest merge rate (78%), vs ~87% for bug fixes.
• Longer prompts correlate with closure (184 vs 160 words), suggesting concise, focused scope helps.
• File references alone are not a differentiator in this period; the overall rate is stable at ~85%.

Recommendations

1. DO: Keep tasks narrowly scoped, one behavior change per PR.
2. DO: Frame features with concrete acceptance criteria to lift the feature merge rate.
3. AVOID: Long, multi-concern bodies and environment-specific changes without reproduction detail.

Historical Trends

Recent history (merge rate among completed) is stable at ~84%: 2026-08-31 84.3% (994 PRs) and 2026-09 entries ~84.3%; today 84.7% (575 PRs).

Trend: Volume dropped (994 → 575 PRs in the 30-day window) while merge rate held steady.

---

Generated by Copilot PR Prompt Analysis (Run: 37989104819)

Generated by 🔍 Copilot PR Prompt Pattern Analysis · copilot · auto · 13.6 AIC · ⌖ 0.689 AIC · ⊞ 10.9K · ◷

• expires on Oct 10, 2026, 12:50 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 #67514.