Every commit scored for change-risk against this repo's own history, so 'elevated' means elevated here rather than on some global curve.
Needs review
11 commits sit in this repo's top risk tercile, which is 33% of the 33scored. The cut is drawn against this codebase's own history rather than a global curve, so a quiet repo still fills its top band, and here it starts at 8.8 out of 10. What pushes a commit up is size and spread together: a large change confined to one area scores below a smaller one scattered across a dozen files.
Commit categories over time, read off the subject line. Fixes carry the accent because that is the series this chart exists to show.
Began other-led, now leaning docs.
Ranked by change-risk, highest first. Priority is a tercile of this repo's own distribution, so a quiet repo still fills its top band.
| # | Commit | Author | When | Lines | Risk | Top driver |
|---|---|---|---|---|---|---|
| 1 | 544c6798CodeSearch 2.0: AI tutor platform across 4 judges, auth, progress, deploy config | Shivam Guptaclaudeassisted | 4w ago | +19.9K -4.5K | 98%Elevated | more lines added than baseline |
| 2 | 06ca040fInitial commit from Create Next App | Shivam Gupta | 4w ago | +7.1K -0 | 95%Elevated | more lines added than baseline |
| 3 | 0d296dcaRedesign tutor workspace: workbench + tutor room | Shivam Gupta | 3w ago | +833 -425 | 89%Elevated | more lines added than baseline |
| 4 | 9114e597Prompts folder, tutor refactor, and hint-leakage eval harness | Shivam Gupta | 3w ago | +1.6K -979 | 89%Elevated | more lines added than baseline |
| 5 | 36583877Interactive learner progress dashboard | Shivam Gupta | 4w ago | +989 -103 | 89%Elevated | more lines added than baseline |
| 6 | b805e02aAdd Sarvam AI vernacular voice tutoring (STT + TTS) | Shivam Gupta | 4w ago | +507 -11 | 83%Elevated | more lines added than baseline |
| 7 | 663af065App-wide dark/light theme with semantic tokens | Shivam Gupta | 3w ago | +409 -269 | 80%Elevated | more lines added than baseline |
| 8 | 894505cfRestore review flow, add practice-next recommendations, thinking indicator | Shivam Gupta | 3w ago | +309 -106 | 77%Elevated | more lines added than baseline |
| 9 | d3f887edAdd demo seed script and Sarvam credits pitch one-pager | Shivam Gupta | 4w ago | +380 -0 | 74%Elevated | more lines added than baseline |
| 10 | 2b51c2c4Vernacular eval, CI workflows, and low-reasoning Sarvam adapter | Shivam Gupta | 3w ago | +427 -5 | 70%Elevated | more lines added than baseline |
| 11 | ac936ca5Add in-browser C++ code runner (Judge0) | Shivam Gupta | 4w ago | +225 -5 | 70%Elevated | more lines added than baseline |
| 12 | 9119303fRewrite README as a product/architecture overview | Shivam Gupta | 4w ago | +195 -22 | 64%Typical | more lines added than baseline |
| 13 | c65fe071Run tutor in-process (CopilotKit + OpenAI) for single-service Vercel deploy | Shivam Guptaclaudeassisted | 4w ago | +211 -168 | 64%Typical | more lines added than baseline |
| 14 | df481528Persist tutor replies and reload chat history | Shivam Gupta | 4w ago | +168 -20 | 59%Typical | more lines added than baseline |
| 15 | ff01b3fcAdd EC2 + Caddy + pm2 deployment recipe (AWS free-tier credits) | Shivam Gupta | 3w ago | +168 -0 | 55%Typical | more lines added than baseline |
| 16 | 50be0f25Sarvam-native tutor brain + spaced-repetition reviews | Shivam Gupta | 4w ago | +233 -14 | 55%Typical | more lines added than baseline |
| 17 | c7cb9461Cohere-style sign-in page in the workbench palette | Shivam Gupta | 3w ago | +160 -32 | 50%Typical | more lines added than baseline |
| 18 | ab48145cRestyle browse and signin pages to the workbench design language | Shivam Gupta | 3w ago | +102 -81 | 45%Typical | more lines added than baseline |
| 19 | d9572ec8Auto-fetch CodeChef statements on first open | Shivam Gupta | 4w ago | +106 -2 | 45%Typical | more lines added than baseline |
| 20 | b8a94ac6Switch code runner to Wandbox; add LeetCode note | Shivam Gupta | 4w ago | +79 -55 | 41%Typical | more lines added than baseline |
| 21 | 8605b5caHarden hint gate, correct eval baselines, gate CI on the T1 invariant | Shivam Gupta | 3w ago | +85 -16 | 38%Typical | more lines added than baseline |
| 22 | 8c87c1eaLet learners pick the tutor model (Sarvam or OpenAI) | Shivam Gupta | 4w ago | +45 -5 | 35%Typical | more lines added than baseline |
| 23 | aba22ce2Prefer own domain in AWS recipe (subdomain split with Vercel) | Shivam Gupta | 3w ago | +9 -6 | 32%Below typical | fewer lines added than baseline |
| 24 | f0cc25b2Guard against undefined visibleMessages in reply capture | Shivam Gupta | 3w ago | +7 -2 | 29%Below typical | fewer lines added than baseline |
| 25 | b37bfb42Personalize AWS recipe to das.boats | Shivam Gupta | 3w ago | +4 -4 | 24%Below typical | fewer lines added than baseline |
| 26 | cef764c9Base difficulty mix on all engaged problems, not just solved | Shivam Gupta | 4w ago | +4 -2 | 24%Below typical | fewer lines added than baseline |
| 27 | 64af6070Log user messages once via onSubmitMessage | Shivam Gupta | 3w ago | +3 -5 | 20%Below typical | fewer lines added than baseline |
| 28 | cb7f6312Dark color-scheme for native form controls on dark pages | Shivam Gupta | 3w ago | +2 -2 | 15%Below typical | fewer lines added than baseline |
| 29 | b4a56325Set maxDuration=60 for streaming tutor replies on Vercel | Shivam Guptaclaudeassisted | 4w ago | +2 -0 | 15%Below typical | fewer lines added than baseline |
| 30 | b776629ePoint docs at das.boats | Shivam Gupta | 3w ago | +2 -2 | 11%Below typical | fewer lines added than baseline |
| 31 | 7fec0615Update README.md | SHIVAM GUPTA | 4w ago | +1 -0 | 8%Below typical | fewer lines added than baseline |
| 32 | 694d882fUpdate README.md | SHIVAM GUPTA | 7h ago | +1 -1 | 3%Below typical | fewer lines added than baseline |
| 33 | 926948fbUpdate route.ts | SHIVAM GUPTA | 23h ago | +1 -1 | 3%Below typical | fewer lines added than baseline |
Two views of the same model: where the cuts fall, and what commit shape lands you above them.
Every scored commit, binned on the raw 0 to 10 score rather than the percentile. Percentile ranks are uniform by construction, so that axis has no shape to draw. The dashed lines are the tercile cuts behind each row's priority pill.
The 33 most recent commits, on their own recency sample rather than the feed above: that defaults to risk-sorted, so reusing it would plot only the top tercile and call it the spread. Big and scattered is what the model penalises. Click a dot to open it.
shivamg5080/codesearch has 33 commits in its history from 2 contributors, the first of them Jun 27, 2026. In the last 90 days 62 files were touched, 117 times in total, most often route.ts. Every commit is scored for change risk from its size, spread and the history of the files it touches.
