What killed your codebase?

Evidence

RECOVERED PROMPT

Rewrite in Rust. Make no mistakes or Grandma dies.

Murder weapon
"quick cleanup"
+18,742  -103
src/app/index.html
RELATED VICTIMS
VibeOS

It's not project management. It's alignment orchestration.

AI-powered clarity for modern teams.

Fingerprint

Git Profile: VibeKing

0 commits 2019-2025, 873 commits in March 2026

Primary suspect

The Vibecoder

  • Adds 12 markdown files to each prompt
  • Recently invested in a Mac Mini
  • Might not be "absolutely right"
Codebase Autopsy
Unused exports
142
Duplicate validators
39
Monolith files
7
API keys committed
4
Alibis
0
Case Method

How We CrackThe Case

A private repo audit runs like a short-lived investigation: get access, secure the scene, split up the leads, file the report.

  1. 01

    The Search Warrant

    You grant us temporary access to your repo via GitHub. We don't need your keys; we just need a look at the files.

  2. 02

    The Stakeout

    Crucial: We clone your codebase into a secure, temporary "interrogation room." We do not keep clone workspaces permanently, and unpublished private scan artifacts are scrubbed after the retention window. We are here to solve the crime, not move in.

  3. 03

    The Lab Work

    Our engine, Brokk, indexes the repo and flags "leads." It then unleashes specialist agents to deep-dive into the mess.

  4. 04

    The Rap Sheet

    We synthesize every lead, duplicate, and dead-end into one final Audit Report. Private unpublished cases are cleaned up automatically after retention; public repository case files remain available unless you delete them or they are removed by support.

Operational Roster

Meet ThePrecinct

  1. Repository Clone

    Det. Checkout

    #1001

    First on the scene. Secures the perimeter, locks down the commit, and preps the case file before the specialists arrive.

  2. Cognitive Complexity

    Det. Knots

    #2199

    Unravels functions whose branching and nesting read like a confession written in mirror writing. Separates real complexity hotspots from harmless structure.

  3. Size & Sprawl

    Det. Sprawl

    #2204

    Tracks down god objects, mile-long methods, and bloated workflow files. Finds the places where one file is doing six jobs and dropping all of them.

  4. Structural Duplication

    Det. Echo

    #3312

    Listens for the telltale echo of duplicated logic: parallel helpers, lookalike abstractions, and homemade mini-frameworks that scatter every future change.

  5. Failure Handling

    Det. Fallback

    #4049

    Investigates suspicious silences. Empty catch blocks, broad excepts, and log-and-continue paths - anywhere a failure was quietly buried instead of handled.

  6. Dead Code & Abstractions

    Det. Morgue

    #3031

    Walks the morgue. Identifies unused declarations, generated-code residue, one-call abstractions, and other deceased citizens still on the books.

  7. Test Signal

    Det. Alibi

    #5172

    Cross-examines the test suite. Flags missing assertions, tautologies, constant-truth checks, and shallow snapshots - the fake alibis that pass without proving a thing.

  8. Comment Intent

    Det. Margins

    #4417

    Reads between the lines. Flags comments that restate syntax, miss the tradeoffs, or paper over the real reason the code looks like that.

  9. Final Report Synthesis

    Lt. Case

    Lt-01

    Reviews every detective's notes, dedupes evidence, runs a final secrets sweep, and signs off on the audit report and scorecard.