VodouAI VODOU QA — Latest Scorecard
Overview
Prints the most recent QA scorecard (.vodou/qa/latest.md, rewritten by scripts/qa/qa.sh on every run) directly into the chat window. The numbers come from the runner, injected pre-LLM — never from the model.
Instructions
⚠️ AUTO-EXECUTE ON LOAD. No greeting, no preamble.
Step 1: Fetch the scorecard (AUTO-EXECUTE)
Call Vodou-script-executor::execute_script with
{"server_name":"vodou-qa","script_name":"qa-report","params":{}}.
Step 2: Render it verbatim
Display the returned markdown exactly as returned — heading, score line, step table, and every failure block with its log path. Do not summarize, re-order, round, or re-word any number. Unescape \n into real newlines so the table renders.
Step 3: Add at most two lines of context
After the scorecard, optionally add up to two lines: how stale it is (compare the stamp to now) and whether tonight's runner (task 50, 03:10) has fired since. Nothing else.
If the script returns empty, say the runner has not produced a scorecard yet and point at ./vodou-core call Vodou-script-executor execute_script '{"server_name":"vodou-qa","script_name":"qa-nightly"}'. Do not invent results.
Notes
- Source of truth:
PLANS/0.6.28/VODOU-QA/ - Scorecard files:
.vodou/qa/latest.md(human) /.vodou/qa/latest.json(machine) - Findings register for dedupe:
PLANS/0.6.28/VODOU-QA/BREAKAGE-BASELINE.md