brycewang-stanford/Auto-Empirical-Research-Skills / skills/33-Galaxy-Dawn-claude-scholar/skills/skill-improver
skill-improver
VerifiedThis skill should be used when the user asks to "apply skill improvements", "update skill from plan", "execute improvement plan", "fix skill issues", "implement skill recommendations", or mentions applying improvements from quality review reports. Reads improvement-plan-{name}.md files generated by skill-quality-reviewer and intelligently merges and executes the suggested changes to improve Claude Skills quality.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 6.9 KB |
| examples/improvement-plan-example.md | markdown | 2.0 KB |
| examples/update-report-example.md | markdown | 1.2 KB |
| references/error-handling.md | markdown | 812 B |
| references/merge-strategies.md | markdown | 14.1 KB |
| references/plan-format.md | markdown | 10.6 KB |
| references/supported-updates.md | markdown | 791 B |
| scripts/backup-skill.shexecutable | script | 1.4 KB |
| scripts/verify-update.shexecutable | script | 6.9 KB |
scripts/verify-update.sh
View on GitHub#!/bin/bash
# verify-update.sh - Verify skill integrity after applying updates
# Part of skill-improver
# Usage: ./verify-update.sh <skill-path>
# Example: ./verify-update.sh ~/.claude/skills/git-workflow
set -euo pipefail
# Check if path provided
if [ $# -eq 0 ]; then
echo "Usage: $0 <skill-path>"
echo "Example: $0 ~/.claude/skills/git-workflow"
exit 1
fi
SKILL_PATH="$1"
SKILL_NAME=$(basename "$SKILL_PATH")
SKILL_FILE="${SKILL_PATH}/SKILL.md"
# Verification results
PASS=0
FAIL=0
WARN=0
echo "================================"
echo "Skill Update Verification"
echo "================================"
echo "Skill: ${SKILL_NAME}"
echo "Path: ${SKILL_PATH}"
echo ""
# Check 1: SKILL.md exists
echo "Check 1: SKILL.md exists"
if [ -f "$SKILL_FILE" ]; then
echo "✅ PASS: SKILL.md found"
((PASS++))
else
echo "❌ FAIL: SKILL.md not found"
((FAIL++))
echo ""
echo "Verification Summary:"
echo " Passed: ${PASS}"
echo " Failed: ${FAIL}"
echo " Warnings: ${WARN}"
exit 1
fi
echo ""
# Check 2: YAML frontmatter valid
echo "Check 2: YAML frontmatter valid"
if command -v yq &> /dev/null; then
# Use yq if available for proper YAML validation
FRONTMAFTER=$(sed -n '/^---$/,/^---$/{ /^---$/d; p; }' "$SKILL_FILE")
if echo "$FRONTMAFTER" | yq eval '.' > /dev/null 2>&1; then
echo "✅ PASS: YAML frontmatter is valid"
((PASS++))
else
echo "❌ FAIL: Invalid YAML frontmatter"
((FAIL++))
fi
else
# Fallback: basic check for name and description fields
if grep -q "^name:" "$SKILL_FILE" && grep -q "^description:" "$SKILL_FILE"; then
echo "✅ PASS: YAML frontmatter has required fields (name, description)"
((PASS++))
else
echo "❌ FAIL: YAML frontmatter missing required fields"
((FAIL++))
fi
fi
echo ""
# Check 3: Directory structure
echo "Check 3: Directory structure"
DIRS_OK=true
if [ -d "${SKILL_PATH}/references" ]; then
echo " ✅ references/ exists"
else
echo " ℹ️ references/ not found (optional)"
((WARN++))
fi
if [ -d "${SKILL_PATH}/examples" ]; then
echo " ✅ examples/ exists"
else
echo " ℹ️ examples/ not found (optional)"
((WARN++))
fi
if [ -d "${SKILL_PATH}/scripts" ]; then
echo " ✅ scripts/ exists"
# Check if scripts are executable (macOS compatible)
SCRIPT_COUNT=$(find "${SKILL_PATH}/scripts" -type f | wc -l | tr -d ' ')
EXEC_COUNT=0
for script in "${SKILL_PATH}/scripts"/*; do
if [ -f "$script" ] && [ -x "$script" ]; then
((EXEC_COUNT++))
fi
done
if [ "$SCRIPT_COUNT" -gt 0 ] && [ "$EXEC_COUNT" -eq "$SCRIPT_COUNT" ]; then
echo " ✅ All scripts are executable"
elif [ "$SCRIPT_COUNT" -gt 0 ]; then
echo " ⚠️ Some scripts are not executable (${EXEC_COUNT}/${SCRIPT_COUNT} executable)"
((WARN++))
fi
else
echo " ℹ️ scripts/ not found (optional)"
fi
if [ "$DIRS_OK" = true ] || [ ! -d "${SKILL_PATH}/references" ] && [ ! -d "${SKILL_PATH}/examples" ]; then
echo "✅ PASS: Directory structure is valid"
((PASS++))
else
echo "⚠️ WARN: Optional directories missing"
fi
echo ""
# Check 4: SKILL.md size (progressive disclosure check)
echo "Check 4: SKILL.md size (progressive disclosure)"
WORD_COUNT=$(wc -w < "$SKILL_FILE" | tr -d ' ')
if [ "$WORD_COUNT" -lt 3000 ]; then
echo "✅ PASS: SKILL.md is ${WORD_COUNT} words (good progressive disclosure)"
((PASS++))
elif [ "$WORD_COUNT" -lt 5000 ]; then
echo "⚠️ WARN: SKILL.md is ${WORD_COUNT} words (could be more concise)"
((WARN++))
else
echo "❌ FAIL: SKILL.md is ${WORD_COUNT} words (too long, should use references/)"
((FAIL++))
fi
echo ""
# Check 5: No broken references (basic check)
echo "Check 5: Referenced files exist"
BROKEN_REFS=0
# Extract references from SKILL.md
# Look for patterns like "See `references/file.md`" or "**`references/pattern.md`**"
REF_FILES=$(grep -oE 'references/[-_a-zA-Z0-9\.]+' "$SKILL_FILE" | sort -u || true)
if [ -n "$REF_FILES" ]; then
for ref in $REF_FILES; do
REF_PATH="${SKILL_PATH}/${ref}"
if [ -f "$REF_PATH" ]; then
echo " ✅ ${ref} exists"
else
echo " ❌ ${ref} referenced but not found"
((BROKEN_REFS++))
fi
done
if [ "$BROKEN_REFS" -eq 0 ]; then
echo "✅ PASS: All referenced files exist"
((PASS++))
else
echo "❌ FAIL: ${BROKEN_REFS} referenced file(s) not found"
((FAIL++))
fi
else
echo "ℹ️ No references found in SKILL.md"
echo "✅ PASS: No broken references"
((PASS++))
fi
echo ""
# Check 6: Writing style (imperative form check)
echo "Check 6: Writing style (imperative form check)"
SECOND_PERSON=$(grep -iE "you should|you need|you can|you must|you'll|your" "$SKILL_FILE" | wc -l | tr -d ' ')
if [ "$SECOND_PERSON" -eq 0 ]; then
echo "✅ PASS: No second-person language found"
((PASS++))
elif [ "$SECOND_PERSON" -lt 5 ]; then
echo "⚠️ WARN: ${SECOND_PERSON} instance(s) of second-person language"
((WARN++))
# Still count as pass for now
((PASS++))
else
echo "❌ FAIL: ${SECOND_PERSON} instances of second-person language"
echo " (should use imperative form)"
((FAIL++))
fi
echo ""
# Check 7: Description quality (basic check)
echo "Check 7: Description quality"
DESCRIPTION=$(grep "^description:" "$SKILL_FILE" | head -1)
if [ -n "$DESCRIPTION" ]; then
# Check for third person
if echo "$DESCRIPTION" | grep -qi "this skill should be used when"; then
echo " ✅ Uses third-person format"
DESC_OK=true
else
echo " ❌ Does not use third-person format"
DESC_OK=false
fi
# Check length (100-300 characters ideal)
DESC_LEN=$(echo "$DESCRIPTION" | cut -d: -f2- | wc -c | tr -d ' ')
if [ "$DESC_LEN" -ge 100 ] && [ "$DESC_LEN" -le 300 ]; then
echo " ✅ Description length is ${DESC_LEN} characters (good)"
elif [ "$DESC_LEN" -lt 100 ]; then
echo " ⚠️ Description is ${DESC_LEN} characters (too short)"
DESC_OK=false
else
echo " ⚠️ Description is ${DESC_LEN} characters (too long)"
DESC_OK=false
fi
if [ "$DESC_OK" = true ]; then
echo "✅ PASS: Description quality is good"
((PASS++))
else
echo "⚠️ WARN: Description could be improved"
((WARN++))
# Still count as pass for basic check
((PASS++))
fi
else
echo "❌ FAIL: No description found in frontmatter"
((FAIL++))
fi
echo ""
echo "================================"
echo "Verification Summary"
echo "================================"
echo " Passed: ${PASS}"
echo " Failed: ${FAIL}"
echo " Warnings: ${WARN}"
echo ""
if [ "$FAIL" -eq 0 ]; then
echo "✅ Verification PASSED"
echo " Skill structure is valid."
exit 0
else
echo "❌ Verification FAILED"
echo " Please review and fix the failed checks."
exit 1
fi
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |