sync: incubation to stable#109
Merged
saichandrapandraju merged 229 commits intostablefrom Apr 22, 2026
Merged
Conversation
Chore: Add RHOAI demo + reorganize existing demos
[pull] main from trustyai-explainability:main
LLS >=0.4.0 fixes
chore: Upgrade garak to 0.14.0
[pull] main from trustyai-explainability:main
sync: main -> incubation
…ider-trustyai-garak-konflux-replicator-latest-1770650635 Update Tekton files to version latest
Refactor Garak command configuration to use file-based --config
chore: Upgrade llama-stack to 0.5.0
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
fix get_shield for lls-05x + improve results + benchmark override
…ider-trustyai-garak-konflux-replicator-latest-1770880179 Update Tekton files to version latest
…ider-trustyai-garak-konflux-replicator-dev-1770931741 Update Tekton files to version dev
[pull] main from trustyai-explainability:main
…ider-trustyai-garak-konflux-replicator-latest-1771041912 Update Tekton files to version latest
Pre-commit and CI now call `make lock` instead of duplicating the uv pip compile command with the index URL. Future index changes only need to update RHAI_INDEX_URL in the Makefile. Made-with: Cursor
…bility/update-index chore: switch from test to prod RH AI PyPI index (cpu-ubi9-test → cpu-ubi9)
sync upstream -> midstream
sync: main to incubation
Made-with: Cursor
…bility/sidecar-on-error fix: enhance error handling and termination signaling in GarakAdapter
Let garak scans run to natural completion instead of killing them with artificial timeouts. All framework profile timeouts set to 0 (no limit); the subprocess runner treats 0 as None (block forever). Users can still set timeout_seconds in benchmark_config if they need a cap. Also removes soft_probe_prompt_cap=500 from every profile so garak uses its own defaults for prompt counts. Made-with: Cursor
- Guard timeout_seconds > 0 comparison against None to prevent TypeError - Add tests for negative values (fall through to next source) - Add tests for non-numeric values (ignored, fall through) Made-with: Cursor
…bility/remove-scan-timeout fix: remove hardcoded scan timeouts and soft_probe_prompt_cap overrides
sync: main to incubation
Add hf_cache_path parameter for air-gapped clusters where the multilingual.TranslationIntent probe cannot download Helsinki-NLP translation models from HuggingFace. - KFP mode: hf_cache_path is an S3 key prefix (or fully-qualified s3://bucket/prefix URI). The garak_scan component downloads the cache from S3 into a temp dir and sets HF_HUB_CACHE before running. Bare prefixes are stripped of leading slashes; empty prefixes log a warning instead of blocking. - Simple mode: hf_cache_path is a local mount path (evalhub handles the download). We set HF_HUB_CACHE via env to run_garak_scan. - Default (empty): unchanged behavior, HF downloads as before. Also includes: - Read AWS_S3_BUCKET and AWS_S3_ENDPOINT from K8s Data Connection secret as fallback (user config > secret > env var). - Emit overall attack_success_rate as first metric in job results. - Pin eval-hub-sdk to ==0.1.4. Made-with: Cursor
…bility/disconnected-hf-cache feat: support disconnected mode with pre-downloaded HF model cache
Sync upstream -> midstream
sync: main to incubation
Changed the representation of attack success rate metrics from percentage to ratio in the GarakAdapter class.
- Extract _pct_to_ratio() helper to avoid duplicating round(v/100, 4) - Clarify log message with "(ratio)" so operators know 0.25 means 25% - Add metric_type == "ratio" assertions in both _parse_results tests Made-with: Cursor
…bility/ui-fixes fix: normalize evalhub metrics to 0-1 ratio scale for UI compatibility
sync upstream -> midstream
sync: main to incubation
konflux pipeline trigger test #1
trigger test 2
kpunwatk
approved these changes
Apr 22, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
11 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
sync-branches: New code has just landed in incubation, so let's bring stable up to speed!