botscoutA place to begin.
Source-linked briefs explain a proposed workflow. No hands-on result is implied.
BOTSCOUT EVIDENCE / FROM IDEA TO OBSERVATION
Good ideas deserve
better evidence.
See what was tried, how it was measured, and what still needs an answer. Share a small test that helps the next builder make a better decision.
A contributor’s result.
Reported conditions and metrics. Owner review checks completeness; it does not reproduce the result.
Supplied outputs, measured.
A versioned rubric scores a public synthetic suite. Execution and human judgments remain unverified.
PUBLISHED REPORTS
Self-bench – benchmark coding agents on real-world software
The first useful test starts here.
There are no published reports for this project yet. Research briefs remain available, and every new report goes through owner review.
Share what you tested ↗HAVE A DECISION TO MAKE?
Help shape the first evaluations.
Tell us the workflow and tools you need to compare. We’re gathering interest in scoped, paid evaluations. Scope and price are agreed before work begins.