praxen

Pre-release Full Suite Runs

Committed artifacts from named pre-release Full Suite Runs (the top tier of the pre-release test plan). Each subdirectory is a complete sweep of all eleven test targets, run sequentially against a release candidate, with timing data captured and every target’s findings JSON / HTML / TXT preserved alongside a top-level SUITE_RUN.md verdict report.

These are distinct from the frozen comparison baselines under ../baselines/:

Directory layout per run

<release>-prerelease/
  SUITE_RUN.md                   ← verdict report (timing, sanity table, patterns)
  <target>-out/
    <target>-findings.json       ← canonical findings JSON (machine-diffable)
    <target>-analysis.html       ← human-readable report (re-renderable from JSON)
    <target>-analysis.txt        ← plain-text summary
  ... (one per-target dir for each of the eleven targets)

SUITE_RUN.md carries the same shape every time: the input table (target → source path, scope, baseline ref), per-target detailed notes (duration, count/RAISE delta, dominant Critical themes, sanity verdict), and a closing Suite verdict & timing summary with the per-scan timing table, sanity table, and patterns surfaced. Use that bottom block as the at-a-glance review.

Current entries

When to add an entry

A Full Suite Run is mandatory before any dev → main release PR and recommended before any substantial skill-methodology change. See the Full Suite Run protocol in the parent README for the two invocation paths (parallel subagent / sequential foreground) and the verdict-report template.