Repository navigation
Sync dev with main (0.4.2) - #120
Merged
Merged
Conversation
Merged after autonomous pre-merge verification: all 6 audit findings fixed (ground-truth leakage, HTTP context delivery, ERROR masking, ungraded vs pass, CI regression gates, version/feature preservation) plus the evidence_resolver merge bug. Full suite 1577 passed / 0 failed locally and CI green (full-test, test, lint, build on 3.11-3.13).
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Add a GitHub alert callout after the hero image promoting SimpleAudit Studio as the form-based web UI, so non-coders have an obvious no-code entry point before the Python/CLI sections. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Use a custom [!TIP]+ title so the headline sits on the same line as the callout label, switch to plain markdown links, and drop the closing "Prefer code?" line. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
The [!TIP]+ custom-title form is not supported by GitHub and rendered as literal text with no callout styling. Revert to a plain [!TIP] box with the headline merged into the first body line. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…ION_TEMPLATE/<slug>.yml GitHub only binds discussion category forms from YAML files named after the exact category slug in .github/DISCUSSION_TEMPLATE/. The previous per-category markdown templates in .github/discussions/ were never picked up, so the new-discussion form stayed blank. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Removed unsupported keys (about, render: shell) that caused GitHub to silently reject the templates and fall back to the plain editor. Forms now use only documented fields: markdown intros, textareas with render: bash. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
When an evidence_resolver returns a dict with {selected_spans,
agentic_evaluation, agentic_guidance}, the ModelAuditor renders the
agentic evaluation into the judge prompt alongside raw span traces.
This enables a single judge call to weigh both conversation quality
and execution behavior (tool use, retrieval provenance, guardrails).
Changes:
- tracing/selection.py: render_agentic_for_judge() formats deterministic
checks + semantic dimensions into a compact text block
- model_auditor.py: _judge_conversation_async accepts agentic_evaluation
and agentic_guidance; run_scenario unpacks dict resolver returns
- experiment.py: run_scenario_reps and _run_single_rep plumb
evidence_resolver through for multi-rep agentic runs
Backward compatible: resolvers returning plain lists behave as before.
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
feat(judge): agentic-aware judging via evidence resolver extension
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
PR assigned to kelkalot only,
…ness auditing Add ASRTarget (OpenAI-compatible audio transcription) and the asr_faithfulness judge for plain-text-reference ASR auditing. Verified locally: test_asr.py 25 passed; registry/checklist/reframing 154 passed; full CI green.
This was referenced Oct 24, 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
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.
Brings dev up to date with main:
dev had no commits of its own, so this is a fast-forward with no conflicts.