chore(digest): flip BRIEF_WHY_MATTERS_SHADOW to opt-in — default-on doubled gemini spend#4909
Merged
Conversation
…oubled gemini spend Shadow mode ran BOTH the analyst and gemini paths on every cache-miss story (Promise.allSettled, one result discarded) and defaulted ON at 100% sampling — a fresh deploy silently paid 2x for whyMatters long after the comparison window ended (#4893). Only exactly '1' enables it now; prod already carries an explicit '0' so behavior there is unchanged either way. Adds a source-text drift pin: the test suite's handwritten readConfig mirror now asserts the handler contains the same opt-in expression, so the mirror can't silently assert stale semantics again. Closes #4893 Claude-Session: https://claude.ai/code/session_01YTm4GsLG2M7ZuaHF9MpZih
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
koala73
enabled auto-merge (squash)
July 5, 2026 17:43
Contributor
Greptile SummaryThis PR flips the
Confidence Score: 5/5Safe to merge — the change is a single expression flip that stops the two-path execution by default, with no impact on the primary response path. The edit is surgically small: one boolean expression changed in the handler, one expression changed in the test mirror, one new test added. The shadow path itself is untouched; only whether it fires by default changes. The existing prod env var (BRIEF_WHY_MATTERS_SHADOW=0) continues to work correctly (still evaluates to false). The new drift-pin test closes the feedback loop that allowed the original regression to go undetected. No files require special attention. Important Files Changed
Flowchart%%{init: {'theme': 'neutral'}}%%
flowchart TD
A[Cache Miss] --> B{cfg.shadowEnabled\nAND sampleHardRoll?}
B -- "BEFORE: default true\n(shadow=ON unless SHADOW=0)" --> C[Promise.allSettled\nanalyst + gemini\n2× LLM spend]
B -- "AFTER: default false\n(shadow=OFF unless SHADOW=1)" --> D{cfg.primary}
D -- analyst --> E[runAnalystPath only\n1× LLM spend]
D -- gemini --> F[runGeminiPath only\n1× LLM spend]
C --> G[choose primary result\nstore shadow diff]
E --> H[cache write + respond]
F --> H
G --> H
%%{init: {'theme': 'base', 'themeVariables': {"darkMode": true, "background": "#0d1117", "primaryColor": "#21262d", "primaryTextColor": "#e6edf3", "primaryBorderColor": "#8b949e", "lineColor": "#8b949e", "textColor": "#e6edf3", "edgeLabelBackground": "#161b22", "actorBkg": "#21262d", "actorBorder": "#8b949e", "actorTextColor": "#e6edf3", "actorLineColor": "#8b949e", "signalColor": "#8b949e", "signalTextColor": "#e6edf3", "noteBkgColor": "#373320", "noteBorderColor": "#d4a72c", "noteTextColor": "#f0e6c0", "labelBoxBkgColor": "#21262d", "labelBoxBorderColor": "#8b949e", "labelTextColor": "#e6edf3", "loopTextColor": "#e6edf3", "activationBkgColor": "#30363d", "activationBorderColor": "#8b949e"}}}%%
flowchart TD
A[Cache Miss] --> B{cfg.shadowEnabled\nAND sampleHardRoll?}
B -- "BEFORE: default true\n(shadow=ON unless SHADOW=0)" --> C[Promise.allSettled\nanalyst + gemini\n2× LLM spend]
B -- "AFTER: default false\n(shadow=OFF unless SHADOW=1)" --> D{cfg.primary}
D -- analyst --> E[runAnalystPath only\n1× LLM spend]
D -- gemini --> F[runGeminiPath only\n1× LLM spend]
C --> G[choose primary result\nstore shadow diff]
E --> H[cache write + respond]
F --> H
G --> H
Reviews (1): Last reviewed commit: "chore(digest): flip BRIEF_WHY_MATTERS_SH..." | Re-trigger Greptile |
This was referenced Jul 5, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Problem (#4893)
api/internal/brief-why-matters.tsshipped shadow mode default-on at 100% sampling: every cache-miss story ranPromise.allSettled([analystPath, geminiPath])— two gemini completions, one used, one stored only as an observability diff. The env flip (BRIEF_WHY_MATTERS_SHADOW=0, already live on Vercel prod) protects the current deployment, but the code default meant any fresh environment silently pays 2× again.Fix
'1'enables (env.BRIEF_WHY_MATTERS_SHADOW === '1'). Prod's explicit0is unchanged/redundant-safe.readConfigmirror now also asserts the handler source contains the same opt-in expression — the mirror asserting stale semantics unnoticed is exactly how default-on shipped in the first place.Tests
tests/brief-why-matters-analyst.test.mjs: default expectation flipped (shadow=OFF), enable-token semantics inverted ('1'enables;yes/0/true/''/ondon't), plus the source-text drift pin. 64/64 ✓,typecheck:api✓, biome ✓.Closes #4893
https://claude.ai/code/session_01YTm4GsLG2M7ZuaHF9MpZih