xfurti/Leogriel behavioral regression

Run paired Agent Skill regressions and publish GitHub-native reports

View on GitHub

Trust Signals

Scorecard Score
not yet scored
Maintenance Recency
Stale
License
None
namedescriptionrequireddefault
skillInstalled project skill name or project-relative skill directoryyes
compareOptional Git ref used as the immutable reference skillno""
agentAgentRunner IDnocodex
runner-versionExact Codex or Claude Code CLI version; latest is convenient but not reproduciblenolatest
modelExact runner model ID; strongly recommended for comparable resultsno""
runsSequential paired runs per test case (1-20)no3
working-directoryProject directory containing agent-skills.jsonno.
leogriel-versionnpm version or dist-tag of @leogriel/clinonext
frozen-installRestore project skills from agent-skills.lock before testingnotrue
trust-testsAllow command assertions from the checked-out test YAMLnofalse
artifact-nameName of the uploaded regression-report artifactnoleogriel-regression
comment-on-prCreate or update the Leogriel pull-request commentnofalse
github-tokenToken used only for an optional pull-request commentno""
namedescription
verdictimproved, unchanged, regressed, or inconclusive
result-jsonAbsolute path to the redacted JSON result
markdown-reportAbsolute path to the Markdown report
html-reportAbsolute path to the HTML report
badge-jsonAbsolute path to a Shields endpoint badge document