Automates iterative bug hunting for the Vole programming language via code generation, testing, reduction, and fixing.
npx skills add https://github.com/pprice/vole --skill stress-hunt
Please help me install this Agent Skill. Skill: stress-hunt Source: https://github.com/pprice/vole/tree/main/.claude/skills/stress-hunt Command: npx skills add https://github.com/pprice/vole --skill stress-hunt
This Skill automates the process of finding and fixing bugs in the Vole programming language by generating random code, testing it, reducing failures, and attempting fixes in an iterative loop.
Invoke the stress-hunt skill to start an automated bug-hunting loop.
High-intent search queries and answers about installing and using this skill.
Automated bug hunting uses an iterative loop to generate random code, test it, reduce failures, and attempt fixes. This skill manages the entire workflow for the Vole compiler, tracking seeds and known bugs in a state file to discover and resolve regressions.
Fuzzing generates random inputs to uncover crashes or unexpected behavior, while differential testing compares outputs across builds. This skill integrates both approaches to find Vole compiler bugs, using AddressSanitizer builds and automated failure reduction to isolate issues.
Automated compiler testing requires generating test programs, executing them, and triaging failures. This skill automates that setup for Vole by running a continuous loop that fingerprints bugs, categorizes them, and dispatches sub-agents to attempt generator or compiler fixes.
Yes, AddressSanitizer builds are supported for automated bug discovery. The skill uses them alongside differential testing to detect memory errors and regressions in the Vole compiler, then minimizes the failing test cases for easier triage.
The best way to triage failing cases is automated failure reduction. This skill fingerprints, categorizes, and minimizes failing Vole programs in its bug hunting loop, preserving the exact conditions needed to reproduce regressions while stripping irrelevant code.
Yes, stress-hunt operates with no external dependencies. It manages its own state file to track round progress, seeds, and known bugs, running the generation, testing, reduction, and fixing workflow entirely self-contained for the Vole programming language.
raphaelchristiConfigure Harness Evolver, LangSmith, and baseline evaluations for Python projects.
dmae97Guide bug resolution from reproduction evidence to root-cause analysis and minimal code changes.
ppriceAdd diagnostic error codes to the Vole compiler with snapshot testing.
ofelixdevAutomates libFuzzer coverage-guided fuzzing of C/C++ codebases with AddressSanitizer and LLVM toolchain instrumentation.
