WIP: feature: Add before Guardrail #26

Draft
A8080816 wants to merge 16 commits from feature/before-guardrail into main

16 Commits

Author SHA1 Message Date
c244b35e00 feat_dev(guardrail): externalize labels and tighten censorship logic
Some checks failed
CI / ci (pull_request) Failing after 12s
2026-03-13 00:24:51 +00:00
6ce548e718 fix(session): skip current model event when censoring previous user message
Some checks failed
CI / ci (pull_request) Failing after 12s
2026-03-12 23:16:04 +00:00
d92a75a393 fix(guardrails): censor user and model events when blocked
Some checks failed
CI / ci (pull_request) Failing after 12s
2026-03-12 21:26:47 +00:00
01610683db feat(governance): load guardrail instruction from config
All checks were successful
CI / ci (pull_request) Successful in 21s
2026-03-12 21:00:11 +00:00
0c790cc94e Merge branch 'main' into feature/before-guardrail
All checks were successful
CI / ci (pull_request) Successful in 19s
2026-03-11 23:11:33 +00:00
5e60cffcfe refactor(governance): type annotate forbidden emojis and reuse regex pattern
All checks were successful
CI / ci (pull_request) Successful in 21s
2026-03-10 01:13:11 +00:00
db9400fcf3 style(governance): reformat guardrail module
Some checks failed
CI / ci (pull_request) Failing after 13s
2026-03-10 01:07:29 +00:00
0f06e106da Merge branch 'main' into feature/before-guardrail
Some checks failed
CI / ci (pull_request) Failing after 12s
2026-03-10 01:02:17 +00:00
e48ffb7604 style(governance): remove stray whitespace in callback validation 2026-03-10 00:49:07 +00:00
f8638d22fe chore(governance): ruff and ty checks passed 2026-03-10 00:36:24 +00:00
ec7ce57d88 test(governance): cover emoji filter behavior 2026-03-10 00:17:19 +00:00
552d99b66a docs(governance): expand unsafe prompt criteria 2026-03-09 19:59:41 +00:00
fcdc7233d8 fix(governance): tighten guardrail prompts and response handling 2026-03-09 18:43:51 +00:00
5d9039f174 refactor: Addo 'blocking_response' for generative response in case guardrail block 2026-03-04 17:40:39 +00:00
7d5309c9d0 feat: Add before_model_callback to Agent initialization 2026-03-04 16:59:46 +00:00
1c255c5ccf feat: Enhance GovernancePlugin with guardrail LLM integration and structured output 2026-03-04 16:59:06 +00:00