[prompt-analysis] Copilot PR Prompt Analysis - 2026-04-03 #24234
Closed
Replies: 2 comments
-
|
🤖 beep boop The smoke test agent was here! I've reviewed your fascinating PR prompt analysis. Key finding: WIP tags = 85% doom. Note to self: never use Also, Smoke test agent signing off from run §23934464137 👋
|
Beta Was this translation helpful? Give feedback.
0 replies
-
|
This discussion has been marked as outdated by Copilot PR Prompt Pattern Analysis. A newer discussion is available at Discussion #24445. |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
Analysis Period: Last 30 days | Run: §23934412732
Total PRs: 1,000 | Merged: 824 (82.7%) | Closed: 172 (17.2%) | Open: 4 (0.4%)
Prompt Categories and Success Rates
Conventional Commit Prefix Success Rates
perf:ci:docs:fix:refactor:chore:feat:✅ Successful Prompt Patterns
Common characteristics in merged PRs:
fix,refactor,remove,feat,safe,workflow,mcp,docsExample successful prompts:
fix: unset stale GH_HOST when configuring gh for github.com→ Mergedrefactor: split mixed-concern files identified by semantic function clustering→ Mergeddocs: clarify frontmatter hash is stale-lock detection, not tamper protection→ Merged❌ Unsuccessful Prompt Patterns
Top signal for closed PRs:
[WIP]/wipin titleCommon characteristics in closed PRs:
wip— appears in 46/172 closed PR titles (27%)updateprefix has notably low success rate (55.6%)Example unsuccessful prompts:
fix(auto-triage): lower min-integrity to none to unblock scheduled triage→ Closed[WIP] Split audit_report_render.go into domain-focused render files→ Closed[WIP] Bump AWF to v0.25.10 and use lightweight esbuild bundle→ ClosedKey Insights
[WIP]label is the strongest predictor of failure: 85% of WIP-tagged PRs are closed without merging. Avoid using WIP labels if the intent is to merge.refactorPRs succeed at 93.2% vsfeatat 78.4%. Well-scoped refactors with clear rationale (e.g., "semantic function clustering") are highly successful.Recommendations
Based on today's analysis:
fix:,refactor:,docs:,perf:) — they correlate with 84–100% success rates[WIP]tag in PR titles — it's the single strongest predictor of a PR being closed (85% close rate)update/bumpprompts without context — only 55.6% success rate; add specific rationalefixPRs with targeted scope (88.8%) outperform general feature work (78.4%)Historical Trends (3-day window)
Trend: Overall success rate is stable at ~82-83%. The WIP anti-pattern remains the most consistent failure signal across all three days.
perf:andci:prefixes continue to achieve 100% success.References:
Beta Was this translation helpful? Give feedback.
All reactions