chris0jeky/llm-release-gate

Block unsafe LLM app changes: run baseline vs candidate configs over a golden dataset, compare quality/abstention/citations/schema/latency/tokens/cost, comment the summary on the PR, and fail the check on a material regression.

View on GitHub

Trust Signals

Scorecard Score
not yet scored
Maintenance Recency
Stale
License
None
namedescriptionrequireddefault
datasetPath to the golden dataset JSONyes
baselinePath to the baseline run-config JSONyes
candidatePath to the candidate run-config JSONyes
scorersPath to the scorer-config JSONyes
thresholdsPath to the thresholds JSONyes
pricingPath to the pricing-table JSON (omit: cost reported unavailable)no""
outOutput directory for reports and the run manifestnollm-release-gate-out
commentPost/update the Markdown summary as a PR comment ('true'/'false')notrue
github-tokenToken used to post the PR commentno${{ github.token }}
python-versionPython version to run the gate withno3.12
namedescription
verdict'pass' or 'fail' (empty if the gate could not run)
result-hashContent hash identifying this comparison result
report-jsonPath to report.json
report-mdPath to report.md
report-htmlPath to report.html