[prompt-analysis] Copilot PR Prompt Analysis - 2026-06-14

[github-actions[bot]]

🤖 Copilot PR Prompt Pattern Analysis — 2026-06-14

Summary

Analysis Period: Last 30 days (2026-05-27 → 2026-06-14)
Total PRs: 1,000 | Merged: 791 (79.1%) | Closed: 202 (20.2%) | Open: 7 (0.7%)

Overall merge rate is stable. Patterns in prompt phrasing, body content, and PR scope show clear correlations with outcomes.

Prompt Categories and Success Rates

Category	Total	Merged	Success Rate
Docs	54	47	87.0%
Test	71	61	85.9%
Chore / Maintenance	234	192	83.1%
Feature Addition	265	204	77.0%
Bug Fix	372	283	76.9%
Refactor	4	4	100% (small sample)

Prompt Analysis

✅ Successful Prompt Patterns

Common characteristics in merged PRs:

• Average body length: 181.6 words (shorter than closed PRs)
• Code blocks present: 84.7% of merged PRs vs 74.3% of closed
• URLs in body: only 13.9% (focused, less link-heavy)
• Highest-success action verbs: update (92.9%), remove (92.3%), docs (90.3%), add (83.9%)

Example successful prompts:

1. PR Narrow cache-memory threat-detection assertion to avoid false positives from AIC guardrail steps #39281: "Narrow cache-memory threat-detection assertion to avoid false positives from AIC guardrail steps" → Merged
2. PR fix(ambient-context): reduce first-request token overhead in smoke-copilot and test-quality-sentinel #39280: "fix(ambient-context): reduce first-request token overhead in smoke-copilot and test-quality-sentinel" → Merged
3. PR Replace raw fmt.Fprintf output in outcomes_history.go with console package #39248: "Replace raw fmt.Fprintf output in outcomes_history.go with console package" → Merged

❌ Unsuccessful Prompt Patterns

Common characteristics in closed PRs:

• Average body length: 211.5 words (longer, often over-scoped)
• Very long bodies (>4000 chars): 5.4% of closed vs only 1.6% of merged
• URLs in body: 22.8% (vs 13.9% for merged — may indicate vague, research-heavy tasks)
• [WIP] in title: only 34% merge rate (18/53 WIP PRs merged)

Example unsuccessful 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{}{}" → Closed
2. PR refactor(workflow): break down 10 largefunc violations in pkg/workflow #39010: "refactor(workflow): break down 10 largefunc violations in pkg/workflow" → Closed
3. PR chore: Extract hard-coded file paths to constants (120 instances) #38774: "chore: Extract hard-coded file paths to constants (120 instances)" → Closed

Key Insights

• Scope over depth: Merged prompts average 30 fewer words than closed ones. Large batch changes ("120 instances", "10 violations") correlate with closure.
• Code blocks signal precision: PRs with code examples merge at 84.7% vs 74.3% without — a concrete body is a strong positive signal.
• WIP is a merge killer: [WIP]-tagged PRs have only a 34% merge rate vs the 79% baseline.

Recommendations

1. DO use specific, scoped verbs: update, remove, add, docs, test — these yield 83–93% success rates.
2. DO include a concrete code example in the PR body — +10 percentage points in merge rate.
3. AVOID open-ended batch refactors and [WIP] tags — both drop merge rates to 34–68%.

Historical Trends (last 7 days)

Date	Merged	Closed	Success Rate
2026-06-14	791	202	79.1%
2026-06-13	792	204	79.5%
2026-06-12	786	206	79.2%
2026-06-10	793	203	79.6%
2026-06-09	787	203	79.5%
2026-06-08	794	199	79.9%
2026-06-07	791	206	79.3%

Trend: Merge rate is stable at ~79.1–79.9%. No significant change detected. Today's rate (79.1%) is within normal variance.

📊 Day-of-Week Success Rates

Day	Merged	Closed	Rate
Sunday	121	21	85.2%
Tuesday	95	18	84.1%
Friday	115	22	83.9%
Saturday	135	30	81.8%
Wednesday	98	29	77.2%
Monday	87	28	75.7%
Thursday	140	54	72.2%

Thursday has the highest volume but lowest rate (72.2%). Sunday–Tuesday average ~84%.

🔑 Branch Keyword Patterns

Deep-dive exploration branches (copilot/deep-*) had 32 merges vs 21 closures — a 60% rate, below the 79% average. Exploratory PRs need tighter scoping.

Verb in title	Merged	Closed	Rate
update	26	2	92.9%
remove	48	4	92.3%
docs	56	6	90.3%
add	151	29	83.9%
fix	129	46	73.7%
refactor	31	14	68.9%
implement	2	2	50.0%

---

Generated by Copilot PR Prompt Analysis — Run: §27511822679

Generated by 🔍 Copilot PR Prompt Pattern Analysis · 430.3 AIC · ⌖ 27.7 AIC · ⊞ 19.6K · ◷

• expires on Jun 15, 2026, 1:11 PM UTC-08:00

[github-actions[bot]]

This discussion was automatically closed because it expired on 2026-06-15T21:11:55.798Z.

Closed by Workflow