Releases: langwatch/scenario
Releases · langwatch/scenario
python: v0.7.24
18 Apr 14:18
Compare
Sorry, something went wrong.
No results found
0.7.24 (2026-04-18)
Features
add GOAT strategy with dynamic technique selection for RedTeamAgent (#306 ) (e62c292 )
python: add async-native scenario.arun for loop-bound resources (#369 ) (a797773 )
python: v0.7.23
10 Apr 11:11
Compare
Sorry, something went wrong.
No results found
0.7.23 (2026-04-08)
Bug Fixes
force verdict on judge discovery exhaustion instead of hard-failing (#315 ) (197f567 )
judge off-by-one, auto-run on script exhaustion, assertion criteria, marathon_script cleanup (#289 ) (91f76d1 )
javascript: v0.4.10
10 Apr 11:12
Compare
Sorry, something went wrong.
No results found
0.4.10 (2026-04-10)
Bug Fixes
default scenarioSetId to 'default' for all events (#305 ) (7bbc8c6 )
default scenarioSetId to "default" when not provided (7bbc8c6 ), closes #304
force verdict on judge discovery exhaustion instead of hard-failing (#315 ) (197f567 )
judge off-by-one, auto-run on script exhaustion, assertion criteria, marathon_script cleanup (#289 ) (91f76d1 )
revert audio model and reduce multilingual test turns (#314 ) (177cdb6 )
revert audio model to gpt-4o-audio-preview and reduce multilingual test turns (177cdb6 )
Miscellaneous
use gpt-5-mini everywhere, enable telemetry, fix reasoning model compat (#311 ) (2384fb2 )
python: v0.7.22
22 Mar 23:56
Compare
Sorry, something went wrong.
No results found
0.7.22 (2026-03-22)
Features
add scenario role and run_id attributes to agent spans (#294 ) (d7e31cc )
javascript: v0.4.9
22 Mar 23:57
Compare
Sorry, something went wrong.
No results found
0.4.9 (2026-03-22)
Features
add scenario role and run_id attributes to agent spans (#294 ) (d7e31cc )
python: v0.7.21
13 Mar 13:04
Compare
Sorry, something went wrong.
No results found
0.7.21 (2026-03-13)
Features
dual conversation histories for RedTeamAgent (#282 ) (fa45876 )
Bug Fixes
javascript: v0.4.8
13 Mar 12:45
Compare
Sorry, something went wrong.
No results found
0.4.8 (2026-03-13)
Features
dual conversation histories for RedTeamAgent (#282 ) (fa45876 )
support optional runId in RunOptions (#284 ) (d5fd769 )
python: v0.7.20
10 Mar 09:26
Compare
Sorry, something went wrong.
No results found
0.7.20 (2026-03-10)
Features
backtracking on hard refusals for RedTeamAgent (#270 ) (62190a0 )
Bug Fixes
align red teaming nomenclature between TypeScript and Python (62190a0 )
resolve CI test failures blocking JS publish (#275 ) (049a13b )
javascript: v0.4.7
10 Mar 09:18
Compare
Sorry, something went wrong.
No results found
0.4.7 (2026-03-10)
Features
add backtracking on hard refusals for RedTeamAgent (TypeScript) (#271 ) (79157cd )
backtracking on hard refusals for RedTeamAgent (#270 ) (62190a0 )
Bug Fixes
align red teaming nomenclature between TypeScript and Python (62190a0 )
resolve CI test failures blocking JS publish (#275 ) (049a13b )
python: v0.7.19
08 Mar 12:25
Compare
Sorry, something went wrong.
No results found
0.7.19 (2026-03-07)
Features
add langwatch.origin="simulation" span attribute (#264 ) (30fbdf0 )
Miscellaneous