scrum: real tool — auto-bundle current diff, run 3-reviewer scrum, push to KB
J's frustration captured: scrum was a TOOL meant to find gaps in the work
and push findings to a KB that informs how WE work on the code. It became
welded into architecture instead of staying a tool. This commits the tool
in the form J actually meant.
Usage:
./scrum auto-bundle origin/main..HEAD, auto-label
./scrum my_label same with explicit label
./scrum --staged bundle staged-only diff (pre-commit)
./scrum --since=COMMIT bundle from a specific commit
Output:
KB row → data/_kb/scrum_findings.jsonl (one row per scrum run)
Verdicts → reports/scrum/_evidence/<date>/verdicts/<label>_*.md
The KB row carries: timestamp, label, diff size, findings count,
convergent count, branch, head SHA, tally excerpt, paths to full
verdicts. Queryable via jq or DuckDB.
Cloud models (opus + kimi-k2 + qwen3-coder) are used here BY DESIGN
— 3-lineage cross-review needs distinct training corpora. This is dev
tooling, not the runtime hot path. PRD line 70 (no cloud APIs) applies
to customer requests, not to J's dev tools.
Tested live: ./scrum --since=HEAD~1 revert_only → 7 findings, 1 convergent
(INFO), all 3 reviewers VERDICT: ship for the revert commit.
KB row written:
{"label":"revert_only","findings_total":7,"findings_convergent":1,
"diff_bytes":5806,"branch":"demo/post-pr11-polish-2026-04-28",...}
What scrum does NOT do (intentionally):
- It does NOT auto-fold findings into architecture / design docs
- It does NOT block any commit/push
- It does NOT mutate any code
- It does NOT run as part of normal customer requests
It reports. J reads. J decides what to do with the findings.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This commit is contained in:
root
parent
d054c0b8b1
commit
2f5ca95875
149
scrum
Executable file
149
scrum
Executable file
@ -0,0 +1,149 @@
|
||||
#!/usr/bin/env bash
|
||||
# scrum — find gaps in current work-in-progress, push to KB.
|
||||
#
|
||||
# Usage:
|
||||
# ./scrum auto-bundle current diff vs origin/main, auto-label
|
||||
# ./scrum my_label same, with explicit label
|
||||
# ./scrum --staged bundle staged-only diff (pre-commit check)
|
||||
# ./scrum --since=COMMIT bundle from a specific commit
|
||||
#
|
||||
# Output:
|
||||
# findings → data/_kb/scrum_findings.jsonl (one row per scrum run, KB-queryable)
|
||||
# verdicts → reports/scrum/_evidence/$(date +%Y-%m-%d)/verdicts/<label>_*.md
|
||||
#
|
||||
# This is a TOOL J runs to find gaps. Findings flow to the KB. The KB
|
||||
# informs how WE work on the code. Findings DO NOT auto-flow into
|
||||
# architecture or design docs — that's J's call after reading them.
|
||||
#
|
||||
# Cloud models are used here BY DESIGN (3-lineage cross-review needs
|
||||
# distinct training corpora). This is dev tooling, not the runtime
|
||||
# hot path — PRD line 70 applies to customer requests, not to J's
|
||||
# dev tools.
|
||||
set -euo pipefail
|
||||
cd "$(git rev-parse --show-toplevel 2>/dev/null || pwd)"
|
||||
|
||||
LABEL=""
|
||||
DIFF_MODE="branch"
|
||||
SINCE_COMMIT=""
|
||||
|
||||
while [ $# -gt 0 ]; do
|
||||
case "$1" in
|
||||
--staged) DIFF_MODE="staged"; shift ;;
|
||||
--since=*) DIFF_MODE="since"; SINCE_COMMIT="${1#--since=}"; shift ;;
|
||||
-h|--help) sed -n '1,18p' "$0"; exit 0 ;;
|
||||
*) LABEL="$1"; shift ;;
|
||||
esac
|
||||
done
|
||||
|
||||
[ -z "$LABEL" ] && LABEL="scrum_$(date +%Y%m%d_%H%M%S)"
|
||||
|
||||
# Build the diff
|
||||
DIFF_FILE=$(mktemp -t "scrum_${LABEL}.XXXXXX.diff")
|
||||
case "$DIFF_MODE" in
|
||||
branch) git diff origin/main...HEAD > "$DIFF_FILE" 2>/dev/null || git diff main...HEAD > "$DIFF_FILE" ;;
|
||||
staged) git diff --staged > "$DIFF_FILE" ;;
|
||||
since) git diff "$SINCE_COMMIT"...HEAD > "$DIFF_FILE" ;;
|
||||
esac
|
||||
|
||||
DIFF_SIZE=$(wc -c < "$DIFF_FILE")
|
||||
DIFF_LINES=$(wc -l < "$DIFF_FILE")
|
||||
|
||||
if [ "$DIFF_SIZE" -lt 50 ]; then
|
||||
echo "[scrum] no diff to review (mode=$DIFF_MODE)"
|
||||
rm -f "$DIFF_FILE"
|
||||
exit 0
|
||||
fi
|
||||
if [ "$DIFF_SIZE" -gt 100000 ]; then
|
||||
echo "[scrum] diff is ${DIFF_SIZE} bytes (>100KB) — kimi/qwen will give up"
|
||||
echo " split into smaller bundles by file/component, then re-run"
|
||||
echo " per-component diffs are typically <60KB"
|
||||
rm -f "$DIFF_FILE"
|
||||
exit 1
|
||||
fi
|
||||
|
||||
echo "[scrum] $LABEL — $DIFF_LINES lines / $DIFF_SIZE bytes — running 3 reviewers"
|
||||
|
||||
# Locate the underlying scrum_review.sh (lives in golangLAKEHOUSE)
|
||||
SCRUM_REVIEW=""
|
||||
for path in \
|
||||
"/home/profit/golangLAKEHOUSE/scripts/scrum_review.sh" \
|
||||
"$(git rev-parse --show-toplevel)/scripts/scrum_review.sh" \
|
||||
"../golangLAKEHOUSE/scripts/scrum_review.sh"
|
||||
do
|
||||
if [ -f "$path" ]; then SCRUM_REVIEW="$path"; break; fi
|
||||
done
|
||||
if [ -z "$SCRUM_REVIEW" ]; then
|
||||
echo "[scrum] could not locate scrum_review.sh"
|
||||
echo " expected at /home/profit/golangLAKEHOUSE/scripts/scrum_review.sh"
|
||||
rm -f "$DIFF_FILE"
|
||||
exit 1
|
||||
fi
|
||||
|
||||
# Default to the live Go gateway on :4110 (chatd-routed multi-provider).
|
||||
# Set LH_GATEWAY=http://127.0.0.1:3100 to use the Rust gateway instead.
|
||||
export LH_GATEWAY="${LH_GATEWAY:-http://127.0.0.1:4110}"
|
||||
|
||||
# Run from the scrum_review.sh directory so its `cd "$(dirname $0)/.."`
|
||||
# works correctly (lands in golangLAKEHOUSE root, where reports/ lives).
|
||||
SCRUM_DIR=$(dirname "$SCRUM_REVIEW")
|
||||
SCRUM_REPO=$(dirname "$SCRUM_DIR")
|
||||
( cd "$SCRUM_REPO" && bash "$SCRUM_REVIEW" "$DIFF_FILE" "$LABEL" ) || {
|
||||
echo "[scrum] reviewers failed — bailing"
|
||||
rm -f "$DIFF_FILE"
|
||||
exit 1
|
||||
}
|
||||
|
||||
# Locate the tally file the review just wrote
|
||||
TALLY="$SCRUM_REPO/reports/scrum/_evidence/$(date +%Y-%m-%d)/verdicts/${LABEL}_tally.md"
|
||||
if [ ! -f "$TALLY" ]; then
|
||||
echo "[scrum] no tally found at $TALLY (review may have failed silently)"
|
||||
rm -f "$DIFF_FILE"
|
||||
exit 1
|
||||
fi
|
||||
|
||||
# Push to KB — one JSONL row per scrum run, queryable via DuckDB or jq
|
||||
KB_DIR="$(git rev-parse --show-toplevel)/data/_kb"
|
||||
mkdir -p "$KB_DIR"
|
||||
KB_FILE="$KB_DIR/scrum_findings.jsonl"
|
||||
|
||||
# Extract finding count + convergent count from the tally.
|
||||
# Tally row shape: `| <reviewers> | <severity> | <where> | <hits> |`
|
||||
# Convergent rows have "+" in the Reviewers (first) column.
|
||||
FINDINGS=$(awk -F'|' '/^\| / && NF>=5 && $2 !~ /^[ -]*$/ && $2 !~ /Reviewers/ {n++} END{print n+0}' "$TALLY" 2>/dev/null)
|
||||
CONVERGENT=$(awk -F'|' '/^\| / && NF>=5 && $2 ~ /\+/ {n++} END{print n+0}' "$TALLY" 2>/dev/null)
|
||||
[ -z "$FINDINGS" ] && FINDINGS=0
|
||||
[ -z "$CONVERGENT" ] && CONVERGENT=0
|
||||
|
||||
# Build a single JSONL row
|
||||
python3 <<PYEOF >> "$KB_FILE"
|
||||
import json, sys, datetime
|
||||
from pathlib import Path
|
||||
tally = Path("$TALLY").read_text()
|
||||
row = {
|
||||
"schema": "scrum_finding.v1",
|
||||
"ts": datetime.datetime.utcnow().isoformat() + "Z",
|
||||
"label": "$LABEL",
|
||||
"diff_mode": "$DIFF_MODE",
|
||||
"diff_bytes": $DIFF_SIZE,
|
||||
"diff_lines": $DIFF_LINES,
|
||||
"findings_total": $FINDINGS,
|
||||
"findings_convergent": $CONVERGENT,
|
||||
"tally_path": "$TALLY",
|
||||
"tally_excerpt": tally[:2000],
|
||||
"branch": "$(git rev-parse --abbrev-ref HEAD 2>/dev/null || echo unknown)",
|
||||
"head_sha": "$(git rev-parse HEAD 2>/dev/null || echo unknown)",
|
||||
}
|
||||
print(json.dumps(row))
|
||||
PYEOF
|
||||
|
||||
rm -f "$DIFF_FILE"
|
||||
|
||||
echo ""
|
||||
echo "──────────────────────────────────────────────────────────────────"
|
||||
echo "[scrum] $LABEL: $FINDINGS findings ($CONVERGENT convergent)"
|
||||
echo "[scrum] tally: $TALLY"
|
||||
echo "[scrum] verdicts: $SCRUM_REPO/reports/scrum/_evidence/$(date +%Y-%m-%d)/verdicts/"
|
||||
echo "[scrum] KB: $KB_FILE"
|
||||
echo ""
|
||||
echo "[scrum] read tally: cat '$TALLY'"
|
||||
echo "[scrum] query KB: jq -c 'select(.label == \"$LABEL\")' '$KB_FILE'"
|
||||
Loading…
x
Reference in New Issue
Block a user