Evolution Advisory

What Changed in Your Codebase

We've detected 7 changes that differ from your project's normal patterns. Each change shows what typically happens versus what we observed this time.

0 of 7 resolved

⚙️

Build Duration

CI / Build

What this means: Build took longer than usual.

Typical:

22.50

This Time:

16,343.0

611.6x above typical range

Trigger: 9c650cca

↘ Returned to baseline

Supporting Evidence

⚠️ Action Required

when CI builds and code changes happen together, change focus and build time tend to move together.

What this means: Builds are taking longer. Slower CI feedback loops reduce developer productivity and delay catching issues. Escalated from Worth Monitoring — multiple correlated patterns converge on the same signal families. Changes span files that don't normally change together, suggesting cross-cutting concerns that may be harder to test. Escalated from Worth Monitoring — multiple correlated patterns converge on the same signal families.

Recommendation: Profile the build pipeline to identify bottlenecks. Check for newly added expensive tests or build steps. Review whether these cross-cutting changes have adequate test coverage.

⚠️ Action Required

when CI builds and code changes happen together, file count and build time tend to move together.

What this means: Commits are touching more files than usual, increasing review burden and risk of regressions. Escalated from Worth Monitoring — multiple correlated patterns converge on the same signal families.

Recommendation: Monitor PR sizes. If this persists, investigate whether large refactors need better decomposition.

⚠️ Action Required

when CI builds and code changes happen together, novelty of file pairings and build time tend to move together.

What this means: Files that don't normally change together are being modified in the same commits. This indicates novel, untested interactions that may introduce bugs. Escalated from Needs Attention — multiple correlated patterns converge on the same signal families.

Recommendation: Investigate which files are being combined unexpectedly. Prioritize testing these changes.

Show 1 more supporting evidence

👁️ Worth Monitoring

when CI builds and dependency changes happen together, dependency count and build time tend to move in opposite directions.

What this means: The dependency count is growing, expanding the supply-chain attack surface and potential for version conflicts.

Recommendation: Audit new dependencies for necessity, maintenance status, and known vulnerabilities.

Drift Investigation Prompt

Development pattern shift detected in CI / Build.\n\nSIGNAL: Build Duration is 611.6x above the typical baseline (observed: 16,343.0, typical: 22.50).\nTRIGGER COMMIT: 9c650cca — \n\nRECENT COMMITS (20 total, showing top 5):\n  b9ec8927 — use chainguard images for minio and tidy up configs

 (2 files)\n  52e2bf63 — Added a new agency webinar (#43000)

## I have read the
[CON (7 files)\n  5a3d12ed — docs: update axiom docs (#42983)

 (1 files)\n  e125ad7f — Chore/deprecate use query state with select part 03 (#42734) (10 files)\n  3f059636 — Joshen/fe 2573 table editor user still wants to run the quer (13 files)\n  ... and 15 more commits\n\nCORRELATED PATTERNS:\n  [CRITICAL] when CI builds and code changes happen together, change focus and build time tend to move together.\n    → Profile the build pipeline to identify bottlenecks. Check for newly added expensive tests or build steps.\n    → Review whether these cross-cutting changes have adequate test coverage.\n  [CRITICAL] when CI builds and code changes happen together, file count and build time tend to move together.\n    → Monitor PR sizes. If this persists, investigate whether large refactors need better decomposition.\n  [CRITICAL] when CI builds and code changes happen together, novelty of file pairings and build time tend to move together.\n    → Investigate which files are being combined unexpectedly. Prioritize testing these changes.\n  [INFO] when CI builds and dependency changes happen together, dependency count and build time tend to move in opposite directions.\n    → Confirm test coverage hasn't decreased alongside faster builds.\n\nINVESTIGATE:\n1. Was this change intentional or did the AI drift from goals?\n2. Review commit 9c650cca — what specifically caused the deviation?\n3. Suggest a course correction (not a bug fix — a realignment).\n\nAFTER FIX:\nRun `evo analyze . --verify` to re-analyze and compare against this run.\nIf the change was intentional, no fix needed — accept it in the report.

Use with: Cursor — paste in chat Claude Code — paste in terminal Copilot — paste in chat panel

After investigation:

AI suggests fixes → apply the changes to your code
Run evo analyze . --verify to re-analyze and compare against this run
If the change was intentional, click Accept above to dismiss it

Show technical details

The run duration for this change was 16,343.0. Historically, similar changes had a value of 22.50 ± 18.00.

📝

Co-change Novelty

Version Control

What this means: Unusually low novelty — changes are more repetitive than normal, touching only well-known file combinations.

Typical:

1.00

This Time:

0.0000

587.2x below typical range

Trigger: 2fcf513a fix: reformat yaml and clarify testing in self-hosted oauth

↘ Returned to baseline

Supporting Evidence

⚠️ Action Required

when CI builds and code changes happen together, novelty of file pairings and build time tend to move together.

What this means: File co-change novelty is unusually low. While familiar patterns reduce risk, this could indicate repetitive or mechanical changes rather than normal development. Escalated from Needs Attention — multiple correlated patterns converge on the same signal families.

Recommendation: Review whether recent commits are routine maintenance or a sign of narrow, repetitive modifications that lack expected variety.

ℹ️ Informational

when deployments and code changes happen together, novelty of file pairings tends to move in opposite directions.

What this means: File co-change novelty is unusually low. While familiar patterns reduce risk, this could indicate repetitive or mechanical changes rather than normal development.

Recommendation: Review whether recent commits are routine maintenance or a sign of narrow, repetitive modifications that lack expected variety.

ℹ️ Informational

when dependency changes and code changes happen together, novelty of file pairings tends to move in opposite directions.

What this means: File co-change novelty is unusually low. While familiar patterns reduce risk, this could indicate repetitive or mechanical changes rather than normal development.

Recommendation: Review whether recent commits are routine maintenance or a sign of narrow, repetitive modifications that lack expected variety.

Drift Investigation Prompt

Development pattern shift detected in Version Control.\n\nSIGNAL: Co-change Novelty is 587.2x below the typical baseline (observed: 0.0000, typical: 1.00).\nTRIGGER COMMIT: 2fcf513a — fix: reformat yaml and clarify testing in self-hosted oauth how-to\n\nFILES CHANGED IN TRIGGER (2):\n  - apps/docs/content/guides/self-hosting/self-hosted-oauth.mdx\n  - apps/docs/content/guides/self-hosting/self-hosted-phone-mfa.mdx\n\nRECENT COMMITS (20 total, showing top 5):\n  b9ec8927 — use chainguard images for minio and tidy up configs

 (2 files)\n  52e2bf63 — Added a new agency webinar (#43000)

## I have read the
[CON (7 files)\n  5a3d12ed — docs: update axiom docs (#42983)

 (1 files)\n  e125ad7f — Chore/deprecate use query state with select part 03 (#42734) (10 files)\n  3f059636 — Joshen/fe 2573 table editor user still wants to run the quer (13 files)\n  ... and 15 more commits\n\nCORRELATED PATTERNS:\n  [CRITICAL] when CI builds and code changes happen together, novelty of file pairings and build time tend to move together.\n    → Investigate which files are being combined unexpectedly. Prioritize testing these changes.\n  [INFO] when deployments and code changes happen together, novelty of file pairings tends to move in opposite directions.\n    → Review whether recent commits are routine maintenance or a sign of narrow, repetitive modifications that lack expected variety.\n  [INFO] when dependency changes and code changes happen together, novelty of file pairings tends to move in opposite directions.\n    → Review whether recent commits are routine maintenance or a sign of narrow, repetitive modifications that lack expected variety.\n\nINVESTIGATE:\n1. Was this change intentional or did the AI drift from goals?\n2. Review commit 2fcf513a — what specifically caused the deviation?\n3. Suggest a course correction (not a bug fix — a realignment).\n\nAFTER FIX:\nRun `evo analyze . --verify` to re-analyze and compare against this run.\nIf the change was intentional, no fix needed — accept it in the report.

Use with: Cursor — paste in chat Claude Code — paste in terminal Copilot — paste in chat panel

After investigation:

AI suggests fixes → apply the changes to your code
Run evo analyze . --verify to re-analyze and compare against this run
If the change was intentional, click Accept above to dismiss it

🚀

Release Cadence

Deployment

What this means: Longer time between releases than usual.

Typical:

745.1

This Time:

25,573.6

229.1x above typical range

Trigger: master

↘ Returned to baseline

Supporting Evidence

⚠️ Action Required

when deployments and code changes happen together, code spread and release frequency tend to move together.

What this means: Changes are spreading across unrelated parts of the codebase. This makes reviews harder and increases the chance of unexpected side effects. Escalated from Worth Monitoring — multiple correlated patterns converge on the same signal families.

Recommendation: Review recent PRs for scope creep. Consider breaking large changes into focused commits.

⚠️ Action Required

when deployments and code changes happen together, file count and release frequency tend to move together.

Recommendation: Monitor PR sizes. If this persists, investigate whether large refactors need better decomposition.

👁️ Worth Monitoring

when dependency changes and deployments happen together, dependency count and release frequency tend to move together.

What this means: The dependency count is growing, expanding the supply-chain attack surface and potential for version conflicts.

Recommendation: Audit new dependencies for necessity, maintenance status, and known vulnerabilities.

Drift Investigation Prompt

Development pattern shift detected in Deployment.\n\nSIGNAL: Release Cadence is 229.1x above the typical baseline (observed: 25,573.6, typical: 745.1).\nTRIGGER COMMIT: master — \n\nRECENT COMMITS (20 total, showing top 5):\n  b9ec8927 — use chainguard images for minio and tidy up configs

 (2 files)\n  52e2bf63 — Added a new agency webinar (#43000)

## I have read the
[CON (7 files)\n  5a3d12ed — docs: update axiom docs (#42983)

 (1 files)\n  e125ad7f — Chore/deprecate use query state with select part 03 (#42734) (10 files)\n  3f059636 — Joshen/fe 2573 table editor user still wants to run the quer (13 files)\n  ... and 15 more commits\n\nCORRELATED PATTERNS:\n  [CRITICAL] when deployments and code changes happen together, code spread and release frequency tend to move together.\n    → Review recent PRs for scope creep. Consider breaking large changes into focused commits.\n  [CRITICAL] when deployments and code changes happen together, file count and release frequency tend to move together.\n    → Monitor PR sizes. If this persists, investigate whether large refactors need better decomposition.\n  [WATCH] when dependency changes and deployments happen together, dependency count and release frequency tend to move together.\n    → Audit new dependencies for necessity, maintenance status, and known vulnerabilities.\n\nINVESTIGATE:\n1. Was this change intentional or did the AI drift from goals?\n2. Review commit master — what specifically caused the deviation?\n3. Suggest a course correction (not a bug fix — a realignment).\n\nAFTER FIX:\nRun `evo analyze . --verify` to re-analyze and compare against this run.\nIf the change was intentional, no fix needed — accept it in the report.

Use with: Cursor — paste in chat Claude Code — paste in terminal Copilot — paste in chat panel

After investigation:

AI suggests fixes → apply the changes to your code
Run evo analyze . --verify to re-analyze and compare against this run
If the change was intentional, click Accept above to dismiss it

Show technical details

The release cadence hours for this change was 25,573.6. Historically, similar changes had a value of 745.1 ± 73.09.

📝

Files Changed

Version Control

What this means: More files changed than usual in this commit.

Typical:

2.00

This Time:

101

66.8x above typical range

Trigger: 156f9040 feat: floating mobile toolbar (#43444)

→ Still elevated (latest deviation: 4.7σ)

Supporting Evidence

⚠️ Action Required

when CI builds and code changes happen together, file count and build time tend to move together.

Recommendation: Monitor PR sizes. If this persists, investigate whether large refactors need better decomposition.

⚠️ Action Required

when deployments and code changes happen together, file count and release frequency tend to move together.

Recommendation: Monitor PR sizes. If this persists, investigate whether large refactors need better decomposition.

👁️ Worth Monitoring

when code changes occur, file count tends to increase.

What this means: Commits are touching more files than usual, increasing review burden and risk of regressions.

Recommendation: Monitor PR sizes. If this persists, investigate whether large refactors need better decomposition.

Drift Investigation Prompt

Development pattern shift detected in Version Control.\n\nSIGNAL: Files Changed is 66.8x above the typical baseline (observed: 101, typical: 2.00).\nTRIGGER COMMIT: 156f9040 — feat: floating mobile toolbar (#43444)\n\nFILES CHANGED IN TRIGGER (101):\n  - apps/studio/components/interfaces/App/CommandMenu/CommandMenu.tsx\n  - apps/studio/components/interfaces/App/FeaturePreview/FeaturePreviewContext.tsx\n  - apps/studio/components/interfaces/App/FeaturePreview/FeaturePreviewModal.tsx\n  - apps/studio/components/interfaces/App/FeaturePreview/FloatingMobileToolbarPreview.tsx\n  - apps/studio/components/interfaces/App/FeaturePreview/useFeaturePreviews.ts\n  - apps/studio/components/interfaces/Home/ProjectList/EmptyStates.tsx\n  - apps/studio/components/interfaces/Organization/ProjectClaim/layout.tsx\n  - apps/studio/components/interfaces/Sidebar.tsx\n  - apps/studio/components/layouts/AccountLayout/AccountLayout.tsx\n  - apps/studio/components/layouts/DefaultLayout.tsx\n  - apps/studio/components/layouts/Navigation/FloatingMobileToolbar/FloatingMobileToolbar.tsx\n  - apps/studio/components/layouts/Navigation/FloatingMobileToolbar/FloatingMobileToolbar.utils.test.ts\n  - apps/studio/components/layouts/Navigation/FloatingMobileToolbar/FloatingMobileToolbar.utils.ts\n  - apps/studio/components/layouts/Navigation/FloatingMobileToolbar/useFloatingToolbarDrag.ts\n  - apps/studio/components/layouts/Navigation/FloatingMobileToolbar/useFloatingToolbarNavSize.ts\n  ... and 86 more\n\nRECENT COMMITS (20 total, showing top 5):\n  b9ec8927 — use chainguard images for minio and tidy up configs

 (2 files)\n  52e2bf63 — Added a new agency webinar (#43000)

## I have read the
[CON (7 files)\n  5a3d12ed — docs: update axiom docs (#42983)

 (1 files)\n  e125ad7f — Chore/deprecate use query state with select part 03 (#42734) (10 files)\n  3f059636 — Joshen/fe 2573 table editor user still wants to run the quer (13 files)\n  ... and 15 more commits\n\nCORRELATED PATTERNS:\n  [CRITICAL] when CI builds and code changes happen together, file count and build time tend to move together.\n    → Monitor PR sizes. If this persists, investigate whether large refactors need better decomposition.\n  [CRITICAL] when deployments and code changes happen together, file count and release frequency tend to move together.\n    → Monitor PR sizes. If this persists, investigate whether large refactors need better decomposition.\n  [WATCH] when code changes occur, file count tends to increase.\n    → Monitor PR sizes. If this persists, investigate whether large refactors need better decomposition.\n\nINVESTIGATE:\n1. Was this change intentional or did the AI drift from goals?\n2. Review commit 156f9040 — what specifically caused the deviation?\n3. Suggest a course correction (not a bug fix — a realignment).\n\nAFTER FIX:\nRun `evo analyze . --verify` to re-analyze and compare against this run.\nIf the change was intentional, no fix needed — accept it in the report.

Use with: Cursor — paste in chat Claude Code — paste in terminal Copilot — paste in chat panel

After investigation:

AI suggests fixes → apply the changes to your code
Run evo analyze . --verify to re-analyze and compare against this run
If the change was intentional, click Accept above to dismiss it

Show technical details

The files touched for this change was 101. Historically, similar changes had a value of 2.00 ± 1.00.

📝

Change Locality

Version Control

What this means: The changed files frequently change together, suggesting a focused modification.

Typical:

0.0000

This Time:

1.00

44.5x above typical range

Trigger: dcbaaad6 fix(common): decode URL-encoded mw_diag cookie before parsin

↘ Returned to baseline

Supporting Evidence

⚠️ Action Required

when CI builds and code changes happen together, change focus and build time tend to move together.

Recommendation: Profile the build pipeline to identify bottlenecks. Check for newly added expensive tests or build steps. Review whether these cross-cutting changes have adequate test coverage.

Drift Investigation Prompt

Development pattern shift detected in Version Control.\n\nSIGNAL: Change Locality is 44.5x above the typical baseline (observed: 1.00, typical: 0.0000).\nTRIGGER COMMIT: dcbaaad6 — fix(common): decode URL-encoded mw_diag cookie before parsing (#43606)\n\nFILES CHANGED IN TRIGGER (2):\n  - packages/common/first-referrer-cookie.test.ts\n  - packages/common/first-referrer-cookie.ts\n\nRECENT COMMITS (20 total, showing top 5):\n  b9ec8927 — use chainguard images for minio and tidy up configs

 (2 files)\n  52e2bf63 — Added a new agency webinar (#43000)

## I have read the
[CON (7 files)\n  5a3d12ed — docs: update axiom docs (#42983)

 (1 files)\n  e125ad7f — Chore/deprecate use query state with select part 03 (#42734) (10 files)\n  3f059636 — Joshen/fe 2573 table editor user still wants to run the quer (13 files)\n  ... and 15 more commits\n\nCORRELATED PATTERNS:\n  [CRITICAL] when CI builds and code changes happen together, change focus and build time tend to move together.\n    → Profile the build pipeline to identify bottlenecks. Check for newly added expensive tests or build steps.\n    → Review whether these cross-cutting changes have adequate test coverage.\n\nINVESTIGATE:\n1. Was this change intentional or did the AI drift from goals?\n2. Review commit dcbaaad6 — what specifically caused the deviation?\n3. Suggest a course correction (not a bug fix — a realignment).\n\nAFTER FIX:\nRun `evo analyze . --verify` to re-analyze and compare against this run.\nIf the change was intentional, no fix needed — accept it in the report.

Use with: Cursor — paste in chat Claude Code — paste in terminal Copilot — paste in chat panel

After investigation:

AI suggests fixes → apply the changes to your code
Run evo analyze . --verify to re-analyze and compare against this run
If the change was intentional, click Accept above to dismiss it

📝

Change Dispersion

Version Control

What this means: Changes spread across unrelated areas of the codebase.

Typical:

0.0000

This Time:

1.05

15.4x above typical range

Trigger: 32330e26 New Campaign Pages: /go (#42920)

↘ Returned to baseline

Supporting Evidence

⚠️ Action Required

when deployments and code changes happen together, code spread and release frequency tend to move together.

Recommendation: Review recent PRs for scope creep. Consider breaking large changes into focused commits.

👁️ Worth Monitoring

when dependency changes and code changes happen together, code spread tends to move together.

What this means: Changes are spreading across unrelated parts of the codebase. This makes reviews harder and increases the chance of unexpected side effects.

Recommendation: Review recent PRs for scope creep. Consider breaking large changes into focused commits.

Drift Investigation Prompt

Development pattern shift detected in Version Control.\n\nSIGNAL: Change Dispersion is 15.4x above the typical baseline (observed: 1.05, typical: 0.0000).\nTRIGGER COMMIT: 32330e26 — New Campaign Pages: /go (#42920)\n\nFILES CHANGED IN TRIGGER (44):\n  - apps/www/README.md\n  - apps/www/_go/index.tsx\n  - apps/www/_go/lead-gen/example-lead-gen.tsx\n  - apps/www/_go/legal/example-legal.tsx\n  - apps/www/_go/thank-you/example-thank-you.tsx\n  - apps/www/app/go/[slug]/page.tsx\n  - apps/www/components/Go/GoPageRenderer.tsx\n  - apps/www/components/Go/TweetsSection.tsx\n  - apps/www/lib/go.ts\n  - apps/www/next.config.mjs\n  - apps/www/package.json\n  - apps/www/pages/state-of-startups.tsx\n  - apps/www/tailwind.config.js\n  - apps/www/types/go.ts\n  - packages/marketing/index.ts\n  ... and 29 more\n\nRECENT COMMITS (20 total, showing top 5):\n  b9ec8927 — use chainguard images for minio and tidy up configs

 (2 files)\n  52e2bf63 — Added a new agency webinar (#43000)

## I have read the
[CON (7 files)\n  5a3d12ed — docs: update axiom docs (#42983)

 (1 files)\n  e125ad7f — Chore/deprecate use query state with select part 03 (#42734) (10 files)\n  3f059636 — Joshen/fe 2573 table editor user still wants to run the quer (13 files)\n  ... and 15 more commits\n\nCORRELATED PATTERNS:\n  [CRITICAL] when deployments and code changes happen together, code spread and release frequency tend to move together.\n    → Review recent PRs for scope creep. Consider breaking large changes into focused commits.\n  [WATCH] when dependency changes and code changes happen together, code spread tends to move together.\n    → Review recent PRs for scope creep. Consider breaking large changes into focused commits.\n\nINVESTIGATE:\n1. Was this change intentional or did the AI drift from goals?\n2. Review commit 32330e26 — what specifically caused the deviation?\n3. Suggest a course correction (not a bug fix — a realignment).\n\nAFTER FIX:\nRun `evo analyze . --verify` to re-analyze and compare against this run.\nIf the change was intentional, no fix needed — accept it in the report.

Use with: Cursor — paste in chat Claude Code — paste in terminal Copilot — paste in chat panel

After investigation:

AI suggests fixes → apply the changes to your code
Run evo analyze . --verify to re-analyze and compare against this run
If the change was intentional, click Accept above to dismiss it

📦

Total Dependencies

Dependencies

What this means: Dependency count increased.

Typical:

2,275.0

This Time:

2,278

4.0x above typical range

Trigger: 7cadc121

↘ Returned to baseline

Supporting Evidence

👁️ Worth Monitoring

when dependency changes and deployments happen together, dependency count and release frequency tend to move together.

What this means: The dependency count is growing, expanding the supply-chain attack surface and potential for version conflicts.

Recommendation: Audit new dependencies for necessity, maintenance status, and known vulnerabilities.

👁️ Worth Monitoring

when CI builds and dependency changes happen together, dependency count and build time tend to move in opposite directions.

What this means: The dependency count is growing, expanding the supply-chain attack surface and potential for version conflicts.

Recommendation: Audit new dependencies for necessity, maintenance status, and known vulnerabilities.

Drift Investigation Prompt

Development pattern shift detected in Dependencies.\n\nSIGNAL: Total Dependencies is 4.0x above the typical baseline (observed: 2,278, typical: 2,275.0).\nTRIGGER COMMIT: 7cadc121 — \n\nRECENT COMMITS (20 total, showing top 5):\n  b9ec8927 — use chainguard images for minio and tidy up configs

 (2 files)\n  52e2bf63 — Added a new agency webinar (#43000)

## I have read the
[CON (7 files)\n  5a3d12ed — docs: update axiom docs (#42983)

 (1 files)\n  e125ad7f — Chore/deprecate use query state with select part 03 (#42734) (10 files)\n  3f059636 — Joshen/fe 2573 table editor user still wants to run the quer (13 files)\n  ... and 15 more commits\n\nCORRELATED PATTERNS:\n  [WATCH] when dependency changes and deployments happen together, dependency count and release frequency tend to move together.\n    → Audit new dependencies for necessity, maintenance status, and known vulnerabilities.\n  [INFO] when CI builds and dependency changes happen together, dependency count and build time tend to move in opposite directions.\n    → Confirm test coverage hasn't decreased alongside faster builds.\n\nINVESTIGATE:\n1. Was this change intentional or did the AI drift from goals?\n2. Review commit 7cadc121 — what specifically caused the deviation?\n3. Suggest a course correction (not a bug fix — a realignment).\n\nAFTER FIX:\nRun `evo analyze . --verify` to re-analyze and compare against this run.\nIf the change was intentional, no fix needed — accept it in the report.

Use with: Cursor — paste in chat Claude Code — paste in terminal Copilot — paste in chat panel

After investigation:

AI suggests fixes → apply the changes to your code
Run evo analyze . --verify to re-analyze and compare against this run
If the change was intentional, click Accept above to dismiss it

Next Steps

Investigate

Copy the prompt below and paste it into your AI assistant (Claude Code, Cursor, Copilot, ChatGPT). It will identify root causes and suggest fixes.

Fix

Apply the suggested changes. If a deviation was intentional, click Accept on its card above instead.

Verify

Run evo analyze . --verify to re-analyze and compare. A verification banner will show which deviations resolved, improved, or persist.

Investigation Prompt

Development drift analysis for supabase/supabase (Mar 06, 2024 at 09:54 PM to Mar 17, 2026 at 07:13 PM). DEVIATIONS FROM BASELINE: - CI / Build: Build Duration — 22.50 -> 16,343.0 - Version Control: Co-change Novelty — 1.00 -> 0.0000 - Deployment: Release Cadence — 745.1 -> 25,573.6 - Version Control: Files Changed — 2.00 -> 101 - Version Control: Change Locality — 0.0000 -> 1.00 - Version Control: Change Dispersion — 0.0000 -> 1.05 - Dependencies: Total Dependencies — 2,275.0 -> 2,278 Click "Show Full Prompt" to see the complete investigation prompt with evidence...

Development drift analysis for supabase/supabase (Mar 06, 2024 at 09:54 PM to Mar 17, 2026 at 07:13 PM).

DEVIATIONS FROM BASELINE:

- CI / Build / Build Duration: 16,343.0 (typical: 22.50, 611.6x above)
- Version Control / Co-change Novelty: 0.0000 (typical: 1.00, 587.2x below)
- Deployment / Release Cadence: 25,573.6 (typical: 745.1, 229.1x above)
- Version Control / Files Changed: 101 (typical: 2.00, 66.8x above)
- Version Control / Change Locality: 1.00 (typical: 0.0000, 44.5x above)
- Version Control / Change Dispersion: 1.05 (typical: 0.0000, 15.4x above)
- Dependencies / Total Dependencies: 2,278 (typical: 2,275.0, 4.0x above)

RISK PATTERNS (actionable only):

- [Action Required] 3 patterns:
    * when deployments and code changes happen together, novelty of file pairings tends to move together.
    * when CI builds and code changes happen together, novelty of file pairings and build time tend to move together.
    * when CI builds and code changes happen together, novelty of file pairings tends to move together.
  Impact: Files that don't normally change together are being modified in the same commits. This indicates novel, untested interactions that may introduce bugs. Escalated from Needs Attention — multiple correlated patterns converge on the same signal families.
  Action: Investigate which files are being combined unexpectedly. Prioritize testing these changes.

- [Action Required] when deployments and code changes happen together, change focus and release frequency tend to move together.
  Impact: Time between releases is increasing. This could indicate a bottleneck in the release process or accumulating risk in larger releases. Escalated from Worth Monitoring — multiple correlated patterns converge on the same signal families.
  Action: Check if process changes or staffing issues are delaying releases.

- [Action Required] when CI builds and code changes happen together, change focus and build time tend to move together.
  (2 independent confirmations)
  Impact: Builds are taking longer. Slower CI feedback loops reduce developer productivity and delay catching issues. Escalated from Worth Monitoring — multiple correlated patterns converge on the same signal families.
  Action: Profile the build pipeline to identify bottlenecks. Check for newly added expensive tests or build steps.

- [Action Required] 4 patterns:
    * when deployments and code changes happen together, code spread and release frequency tend to move together.
    * when CI builds and code changes happen together, code spread and build time tend to move together.
    * when deployments and code changes happen together, code spread tends to move together.
    * when CI builds and code changes happen together, code spread tends to move together.
  Impact: Changes are spreading across unrelated parts of the codebase. This makes reviews harder and increases the chance of unexpected side effects. Escalated from Worth Monitoring — multiple correlated patterns converge on the same signal families.
  Action: Review recent PRs for scope creep. Consider breaking large changes into focused commits.

- [Action Required] 4 patterns:
    * when deployments and code changes happen together, file count tends to move together.
    * when CI builds and code changes happen together, file count and build time tend to move together.
    * when deployments and code changes happen together, file count and release frequency tend to move together.
    * when CI builds and code changes happen together, file count tends to move together.
  Impact: Commits are touching more files than usual, increasing review burden and risk of regressions. Escalated from Worth Monitoring — multiple correlated patterns converge on the same signal families.
  Action: Monitor PR sizes. If this persists, investigate whether large refactors need better decomposition.

- [Action Required] when CI builds and code changes happen together, change focus tends to move in opposite directions.
  Impact: Changes span files that don't normally change together, suggesting cross-cutting concerns that may be harder to test. Escalated from Worth Monitoring — multiple correlated patterns converge on the same signal families.
  Action: Review whether these cross-cutting changes have adequate test coverage.

- [Worth Monitoring] when dependency changes and deployments happen together, dependency count and release frequency tend to move together.
  (2 independent confirmations)
  Impact: The dependency count is growing, expanding the supply-chain attack surface and potential for version conflicts.
  Action: Audit new dependencies for necessity, maintenance status, and known vulnerabilities.

- [Worth Monitoring] when dependency changes and code changes happen together, change focus tends to move in opposite directions.
  Impact: Changes span files that don't normally change together, suggesting cross-cutting concerns that may be harder to test.
  Action: Review whether these cross-cutting changes have adequate test coverage.

- [Worth Monitoring] when dependency changes and code changes happen together, code spread tends to move together.
  Impact: Changes are spreading across unrelated parts of the codebase. This makes reviews harder and increases the chance of unexpected side effects.
  Action: Review recent PRs for scope creep. Consider breaking large changes into focused commits.

- [Worth Monitoring] when code changes occur, file count tends to increase.
  Impact: Commits are touching more files than usual, increasing review burden and risk of regressions.
  Action: Monitor PR sizes. If this persists, investigate whether large refactors need better decomposition.


COMMITS (20):

  b9ec8927 — use chainguard images for minio and tidy up configs (2 files)
  52e2bf63 — Added a new agency webinar (#43000) (7 files)
  5a3d12ed — docs: update axiom docs (#42983) (1 files)
  e125ad7f — Chore/deprecate use query state with select part 03 (#42734) (10 files)
  3f059636 — Joshen/fe 2573 table editor user still wants to run the query if it causing (#43 (13 files)
  86f62a2f — docs: update js sdk docs (2.97.0) (#42988) (14 files)
  a36c50ad — add logarithmic option to custom reports (#43001) (7 files)
  2674009c — docs: fix grammar, spelling, and capitalization across docs and studio (#42917) (8 files)
  647ac43c — fix: correct indefinite article usage for acronyms (an HTTP, an API, an RLS) (#4 (13 files)
  682931e6 — fix: Bump resizable package and fix minor issues (#42892) (5 files)
  ... and 10 more

SOURCE FILES CHANGED (50):

  - docker/docker-compose.s3.yml (modified)
  - docker/docker-compose.yml (modified)
  - apps/docs/content/guides/database/joins-and-nesting.mdx (modified)
  - apps/docs/content/guides/platform/read-replicas/getting-started.mdx (modified)
  - apps/docs/content/guides/self-hosting/copy-from-platform-s3.mdx (modified)
  - apps/docs/content/guides/self-hosting/self-hosted-s3.mdx (modified)
  - apps/docs/content/guides/telemetry/log-drains.mdx (modified)
  - apps/www/_events/2026-03-19-agency-webinar-ai-prototyping-production.mdx (modified)
  - apps/www/public/rss.xml (modified)
  - apps/studio/components/interfaces/Database/EnumeratedTypes/DeleteEnumeratedTypeModal.tsx (modified)
  - apps/studio/components/interfaces/Database/EnumeratedTypes/EnumeratedTypes.tsx (modified)
  - apps/studio/components/interfaces/Database/Functions/DeleteFunction.tsx (modified)
  - apps/studio/components/interfaces/Database/Functions/FunctionsList/FunctionsList.tsx (modified)
  - apps/studio/components/interfaces/Database/Roles/DeleteRoleModal.tsx (modified)
  - apps/studio/components/interfaces/Database/Roles/RolesList.tsx (modified)
  - apps/studio/components/interfaces/Database/Triggers/TriggersList/TriggersList.tsx (modified)
  - apps/studio/components/ui/SparkBar.tsx (modified)
  - apps/studio/pages/project/[ref]/database/functions.tsx (modified)
  - e2e/studio/features/database.spec.ts (modified)
  - apps/studio/components/grid/SupabaseGrid.tsx (modified)
  ... and 30 more

TASKS:

1. ROOT CAUSE: For each deviation, identify the commit(s) that caused it.
   Focus on [Action Required] and [Needs Attention] items first.

2. FIXES: Provide concrete fixes with file paths and code changes.
   Goal: bring metrics back toward baseline without disrupting velocity.

3. PRIORITY: Rank fixes by urgency (immediate vs. next sprint).

4. AFTER FIXING: Run `evo analyze . --verify` to confirm deviations decreased.
   If a change was intentional, accept it: `evo accept . <N>`.

5. FINDING SUMMARIES: At the end, include a section like this:
   ## Finding Summaries
   - [family/metric]: One plain-English sentence for a non-technical reader.
   (This lets the user run `evo enrich . --from response.txt` to store friendly descriptions.)

Executive Summary

What EE Can See

Key Findings

What Changed in Your Codebase

Next Steps

Investigation Prompt

Expand Your Coverage

Available Adapters

How to Connect an Adapter