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
8 commits sit in this repo's top risk tercile, which is 36% of the 22scored. 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 8.5 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 feature.
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 | Review priority | Top driver |
|---|---|---|---|---|---|---|
| 1 | 48e6dfdcfix(tightens prompts and tool calling): restricting model to use powerful commands and giving it web search to get latest dependencies | Yash Vardhan Shukla | 2mo ago | +1.6K -339 | 98thElevated | more lines added than baseline |
| 2 | 377605b5graph executing | Yash Vardhan Shukla | 2mo ago | +957 -0 | 93thElevated | more lines added than baseline |
| 3 | d7a6e576fix(path and cli) fixed relative path issue, and approval ui bug | Yash Vardhan Shukla | 2mo ago | +448 -130 | 89thElevated | more lines added than baseline |
| 4 | 4dd683fdfeat(code): agent is now creating projects in single prompt | Yash Vardhan Shukla | 2mo ago | +295 -11 | 82thElevated | more lines added than baseline |
| 5 | c22b03ddcreated fastapi server with /analyze endpoint which calls the agent pipeline | Yash Vardhan Shukla | 2mo ago | +271 -228 | 82thElevated | more lines added than baseline |
| 6 | 126ee862feat(coder): added models and new agents for code writing | Yash Vardhan Shukla | 2mo ago | +302 -41 | 75thElevated | more lines added than baseline |
| 7 | b42e3886fix(architecture): planner dont call all three agents, it only call those which are needed. | Yash Vardhan Shukla | 2mo ago | +213 -4 | 68thElevated | more lines added than baseline |
| 8 | 925c047dall agents set to gemini | Yash Vardhan Shukla | 2mo ago | +217 -4 | 68thElevated | more lines added than baseline |
| 9 | e3abd6bfreadme updated | Yash Vardhan Shukla | 2mo ago | +139 -34 | 61thTypical | more lines added than baseline |
| 10 | 7eb339a3feat(dockerfile) : added dockerfile for app | Yash Vardhan Shukla | 2mo ago | +175 -5 | 57thTypical | more lines added than baseline |
| 11 | 96c78c23feat(mcp): add MCP server wrapping CodeGuard CLI as analyze_codebase tool | Yash Vardhan Shukla | 2mo ago | +90 -21 | 50thTypical | more lines added than baseline |
| 12 | e69fde10add(banner) readme | Yash Vardhan Shukla | 2mo ago | +122 -31 | 50thTypical | more lines added than baseline |
| 13 | da3a47c3Updated Readme | Yash Vardhan Shukla | 2mo ago | +71 -22 | 43thTypical | more lines added than baseline |
| 14 | 8e93788afix(removed review agent) : token usage reduced by 3x | Yash Vardhan Shukla | 2mo ago | +62 -47 | 39thTypical | more lines added than baseline |
| 15 | d863f897fix(strict prompt): restricting model to use commands like cat or ls on large files like node_modules and other big files | Yash Vardhan Shukla | 2mo ago | +42 -5 | 34thTypical | more lines added than baseline |
| 16 | bea2b1defix(str parsing) changed input_data state , so memory is initialised as [] | Yash Vardhan Shukla | 2mo ago | +39 -42 | 30thBelow typical | more lines added than baseline |
| 17 | b4830bf7README updated | Yash Vardhan Shukla | 2mo ago | +26 -8 | 25thBelow typical | more lines added than baseline |
| 18 | b1e053bafeat(fallback) : on rate limit gemini will fallback to qwen | Yash Vardhan Shukla | 2mo ago | +12 -2 | 20thBelow typical | more focused than baseline |
| 19 | b7ebee0dfeat(tracers) : langsmith integrated for tracing | Yash Vardhan Shukla | 2mo ago | +7 -3 | 16thBelow typical | fewer lines added than baseline |
| 20 | ebfb15cfdocs(env) : updated .env example | Yash Vardhan Shukla | 2mo ago | +6 -3 | 11thBelow typical | fewer lines added than baseline |
| 21 | 1044ef91fix(model updated) : using 2.5 flash and removed extra prompt from code_planner | Yash Vardhan Shukla | 2mo ago | +3 -8 | 7thBelow typical | fewer lines added than baseline |
| 22 | 95a3a245(feat): added chat node in the graph, and set all models to gemini-3.1-flash-lite | Yash Vardhan Shukla | 2mo ago | +0 -73 | 2thBelow 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 22 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.
yash-vs9/codeguard has 22 commits in its history from 1 contributor, the first of them Jun 6, 2026. In the last 90 days 28 files were touched, 69 times in total, most often code.py. Every commit is scored for change risk from its size, spread and the history of the files it touches.
