Running Fuzzers

Run fuzz targets on their own, or after your normal test pass.

ast fuzz                 # run every fuzz target
ast test --fuzz          # run specs first, then fuzzers

Useful flags

ast fuzz

Flag	Effect
--runs <value>	Override the iteration count (see formats below).
--seed <n>	Pin the base seed for a reproducible campaign.
--fuzzer <name[,name...]>	Run only matching targets (alias: --fuzzers, --suite).
--mode <name[,name...]>	Run under specific modes.
--jobs <n>	Spread targets across a worker pool.
--list	Preview the resolved fuzz files without running.

ast test --fuzz

Flag	Effect
--fuzz	Run fuzz targets after the test pass.
--fuzz-runs <value>	Override the iteration count for this run.
--fuzz-seed <n>	Pin the base seed for this run.

Run count formats

--runs and --fuzz-runs accept an absolute count or an adjustment relative to the configured fuzz.runs:

Value	Meaning
500	Exactly 500 iterations.
1.5x	1.5× the configured count.
+10%	10% more than configured.
+100000	100,000 more than configured.

ast fuzz --runs 5000
ast fuzz --runs 2x --seed 1337
ast test --fuzz --fuzz-runs +50%

Filtering targets

ast fuzz parser              # files matching "parser"
ast fuzz --fuzzer "addition" # targets whose name matches

Notes

• Fuzz builds require the bindings target (fuzz.target is "bindings"). This is independent of the target your specs use.
• If no fuzz.seed is configured and you don't pass --seed, each campaign picks a random base seed — printed in the output so you can pin it later.
• Configure counts, input size, corpus, and crash output under the fuzz config block.