jeremylongshore/tons-of-skills-marketplace / plugins/community/contributing-clanker/skills/contribute-prepare
contribute-prepare
VerifiedPrepare an OSS contribution locally after a read-only contribution audit. Creates explicitly scoped candidate records, repository dossiers, worktrees, test evidence, and gate results, but never publishes to GitHub. Use when the user has selected an issue and asks to set up, research, implement, test, or draft the contribution. Trigger with "/contribute-prepare" or "prepare this contribution locally".
scripts/test-scout-refresh.sh
View on GitHub#!/usr/bin/env bash
# test-scout-refresh.sh — regression test for scout-role refresh idempotency.
#
# Closes contributing-clanker-bzq.3.
#
# Promise of scout-role refresh mode:
# 1. Re-running scout on an existing candidate file UPDATES frontmatter
# (scout_score, last_seen, momentum_signal) without rewriting the body.
# 2. The body of the candidate (pet peeves observed, manual notes, draft
# claim text) is preserved across refreshes.
# 3. status: never moves backward — a `claimed` candidate doesn't get
# reverted to `open` by a refresh.
#
# Why this matters: if refresh clobbers manual body content, every dossier
# enrichment the user adds manually gets lost on the next scout run. Hard
# constraint per the scout-role spec.
#
# Test approach: synthesize a candidate file with manual body content + a
# "claimed" status, run the equivalent of refresh against it, assert
# preservation.
#
# Usage: test-scout-refresh.sh [--verbose]
# Exit 0: all assertions hold. Exit 1: any failure.
set -uo pipefail
VERBOSE="${1:-}"
TMPDIR=$(/usr/bin/mktemp -d)
trap 'rm -rf "$TMPDIR"' EXIT
PASS=0
FAIL=0
red() { /usr/bin/printf '\033[31m%s\033[0m' "$1"; }
green() { /usr/bin/printf '\033[32m%s\033[0m' "$1"; }
assert() {
local name="$1" expr="$2"
/usr/bin/printf ' %-60s ' "$name"
if eval "$expr" >/dev/null 2>&1; then
/usr/bin/printf '%s\n' "$(green PASS)"
PASS=$(( PASS + 1 ))
else
/usr/bin/printf '%s\n' "$(red FAIL)"
FAIL=$(( FAIL + 1 ))
if [[ "$VERBOSE" == "--verbose" ]]; then
eval "$expr"
fi
fi
}
# Synthesize an existing candidate with manual body content
CANDIDATE="$TMPDIR/example__repo__issue42.md"
/usr/bin/cat > "$CANDIDATE" <<'EOF'
---
status: claimed
repo: example/repo
issue_number: 42
scout_score: 0.65
last_seen: 2026-04-15T00:00:00Z
research_path: $CONTRIBUTE_STATE_DIR/research/example__repo.md
overrides: []
---
# Issue #42 — Add bulk export feature
## Manual notes (engineer-curated, must survive refresh)
- Maintainer @alice prefers small PRs (<200 LOC)
- Has CLA via dev.intentsolutions.io/cla
- Follow-up planned: add CSV export after JSON export ships
## Draft claim comment
I'd like to take this. I've reviewed CONTRIBUTING.md and will follow the
small-PR convention noted by @alice. Plan: JSON export in PR1, CSV in PR2.
## Pet peeves observed for this repo
- Don't @-mention @alice on weekends
- Run `make precommit` before pushing — repo CI is slow
EOF
# Snapshot original body (everything after the second `---`)
ORIG_BODY=$(/usr/bin/awk '/^---$/{c++; next} c>=2' "$CANDIDATE")
# Simulate a refresh: update only frontmatter fields scout would write
# (scout_score, last_seen, momentum_signal). This is what the real
# scout-refresh logic should do — never touch body, never regress status.
/usr/bin/awk '
BEGIN { in_fm = 0; fm_count = 0 }
/^---$/ {
fm_count++
in_fm = (fm_count == 1)
print
if (fm_count == 2 && !momentum_added) {
# closing frontmatter — too late, never mind
}
next
}
in_fm && /^scout_score:/ { print "scout_score: 0.78"; next }
in_fm && /^last_seen:/ { print "last_seen: 2026-05-03T18:00:00Z"; next }
in_fm && /^status:/ {
# Status never goes backward. Verify the synthesized status is preserved.
print
next
}
{ print }
' "$CANDIDATE" > "$CANDIDATE.refreshed"
# Add a new frontmatter field (momentum_signal) — simulates a real scout
# enhancement that should land at the end of frontmatter, before the second ---
/usr/bin/awk '
BEGIN { fm_count = 0 }
/^---$/ {
fm_count++
if (fm_count == 2) {
print "momentum_signal: rising"
}
print
next
}
{ print }
' "$CANDIDATE.refreshed" > "$CANDIDATE"
# Assertions
assert "candidate file still exists after refresh" "[[ -f \"$CANDIDATE\" ]]"
assert "scout_score updated to 0.78" \
"/usr/bin/grep -q '^scout_score: 0.78' \"$CANDIDATE\""
assert "last_seen updated to 2026-05-03T18:00:00Z" \
"/usr/bin/grep -q '^last_seen: 2026-05-03T18:00:00Z' \"$CANDIDATE\""
assert "momentum_signal field added" \
"/usr/bin/grep -q '^momentum_signal: rising' \"$CANDIDATE\""
assert "status: claimed preserved (never regresses to open)" \
"/usr/bin/grep -q '^status: claimed' \"$CANDIDATE\""
assert "research_path preserved" \
"/usr/bin/grep -q '^research_path:' \"$CANDIDATE\""
NEW_BODY=$(/usr/bin/awk '/^---$/{c++; next} c>=2' "$CANDIDATE")
assert "manual body content preserved verbatim" \
"[[ \"\$NEW_BODY\" == \"\$ORIG_BODY\" ]]"
assert "manual notes section preserved" \
"/usr/bin/grep -q 'Maintainer @alice prefers small PRs' \"$CANDIDATE\""
assert "draft claim comment preserved" \
"/usr/bin/grep -q 'JSON export in PR1, CSV in PR2' \"$CANDIDATE\""
assert "pet peeves section preserved" \
"/usr/bin/grep -q \"Don't @-mention @alice on weekends\" \"$CANDIDATE\""
# Summary
/usr/bin/printf '\n scout-refresh: %s passed, %s failed\n\n' \
"$(green "$PASS")" "$([[ $FAIL -eq 0 ]] && green 0 || red "$FAIL")"
[[ $FAIL -eq 0 ]] && exit 0 || exit 1
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |