[prompt-analysis] Copilot PR Prompt Analysis - 2026-07-26

[github-actions[bot]]

🤖 Copilot PR Prompt Pattern Analysis - 2026-07-26

Summary

Analysis Period: Last 30 days
Total PRs: 1000 | Merged: 764 (76.4%) | Closed: 230 (23.0%) | Open: 6 (0.6%)

Prompt Categories and Success Rates

Category	Total	Merged	Success Rate
Documentation	57	49	86%
Testing	41	35	85%
Refactoring	50	40	82%
Bug Fix	619	472	77%
Feature Addition	166	126	76%
Other/Uncategorized	67	42	63%

Prompt Analysis

✅ Successful Prompt Patterns

Common characteristics in merged PRs:

• Average prompt length: 211 words
• Most relevant keywords: workflow, run, actions, copilot, model
• Top action verbs: fix, add, implement, resolve, align

Example Merged PRs

1. PR Add evidence-driven linter workflow guidance #48226: Add evidence-driven linter workflow guidance → Merged
2. PR Propagate models.providers pricing into AWF apiProxy (including threat-detection runs) #48107: Propagate models.providers pricing into AWF apiProxy (including threat-detection) → Merged
3. PR Align repo-memory max-patch-size pre-check with push diff semantics #48106: Align repo-memory max-patch-size pre-check with push diff semantics → Merged

❌ Unsuccessful Prompt Patterns

Common characteristics in closed PRs:

• Average prompt length: 217 words (slightly longer — may indicate over-specification or scope creep)
• Most common keywords: workflow, copilot, daily, agent, runs
• Issues identified: WIP items, large cross-cutting changes, ambiguous scope

Example Closed PRs

1. PR [WIP] Fix safeoutputs gateway registering zero tools issue #48186: [WIP] Fix safeoutputs gateway registering zero tools issue → Closed
2. PR Add Gemini Vertex AI auth via engine.auth #48170: Add Gemini Vertex AI auth via engine.auth → Closed
3. PR Restore embedded cli-proxy digest pin for historical AWF 0.27.11 #48103: Restore embedded cli-proxy digest pin for historical AWF 0.27.11 → Closed

Key Insights

• Documentation and test prompts have highest success rates (85–86%): Prompts focused on docs or test coverage are well-scoped and reviewable, leading to higher merge rates.
• WIP markers correlate with closure: PRs tagged [WIP] or with very broad feature additions (e.g., new auth provider integrations) show higher closure rates — Copilot may not have enough context to complete these fully.
• Prompt length has minimal impact: Merged and closed prompts are nearly identical in length (~211 vs 217 words), suggesting length is not a primary success factor — specificity and scope matter more.

Recommendations

Based on today's analysis:

1. DO: Frame prompts around docs, tests, or focused refactors — these categories consistently show 80%+ success rates.
2. DO: Reference specific files, functions, or error messages in the prompt to increase specificity and reduce ambiguity.
3. AVOID: Using [WIP] in PR titles or submitting broad feature additions spanning multiple components — these have the highest closure rate (37% for "other/uncategorized").

Historical Trends (Last 7 Days)

Date	PRs	Merged	Success Rate
2026-07-19	1000	762	76.2%
2026-07-20	1000	755	75.5%
2026-07-21	1000	764	76.4%
2026-07-22	1000	760	76.0%
2026-07-23	1000	764	76.4%
2026-07-24	1000	766	76.6%
2026-07-26	1000	764	76.4%

Trend: Success rates are stable at 76–77% over the past week with minimal variance. No significant improvement or degradation detected.

---

References:

• §30219990551

Generated by 🔍 Copilot PR Prompt Pattern Analysis · sonnet46 · 21.3 AIC · ⌖ 7.78 AIC · ⊞ 7.6K · ◷

• expires on Jul 27, 2026, 1:01 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 #48449.