返回 Skill 列表
extension
分类: 开发与工程无需 API Key

blueprint-docs-list

列出蓝图文档(ADRs、PRDs、PRPs),并从frontmatter和标题中提取元数据

person作者: jakexiaohubgithub

List blueprint documents programmatically from the filesystem. Extracts metadata from YAML frontmatter and markdown headers.

Use Case: Audit document status, generate index tables, or get a quick overview of all project documentation.

Parameters

| Arg | Description | |-----|-------------| | adrs | List Architecture Decision Records | | prds | List Product Requirements Documents | | prps | List Product Requirement Prompts | | all | Summary of all document types |

Execution

If arg is adrs

Run /blueprint:adr-list — it handles ADR-specific extraction with both header-section and frontmatter support.

If arg is prds

printf "| PRD | Title | Status | Date |\n|-----|-------|--------|------|\n" && \
for f in docs/prds/*.md; do
  [ -f "$f" ] || continue
  fname=$(basename "$f")
  [ "$fname" = "README.md" ] && continue
  doc_title=$(head -50 "$f" | grep -m1 "^title:" | sed 's/^title:[[:space:]]*//' || true)
  doc_status=$(head -50 "$f" | grep -m1 "^status:" | sed 's/^status:[[:space:]]*//' || true)
  doc_date=$(head -50 "$f" | grep -m1 "^date:\|^created:" | sed 's/^[^:]*:[[:space:]]*//' || true)
  # Fallback: extract title from H1
  if [ -z "$doc_title" ]; then
    doc_title=$(head -20 "$f" | grep -m1 "^# " | sed 's/^# //')
  fi
  printf "| [%s](%s) | %s | %s | %s |\n" \
    "${fname%.md}" "$f" "${doc_title:-(untitled)}" "${doc_status:--}" "${doc_date:--}"
done

If arg is prps

printf "| PRP | Title | Status | Confidence |\n|-----|-------|--------|------------|\n" && \
for f in docs/prps/*.md; do
  [ -f "$f" ] || continue
  fname=$(basename "$f")
  [ "$fname" = "README.md" ] && continue
  doc_title=$(head -50 "$f" | grep -m1 "^title:" | sed 's/^title:[[:space:]]*//' || true)
  doc_status=$(head -50 "$f" | grep -m1 "^status:" | sed 's/^status:[[:space:]]*//' || true)
  doc_confidence=$(head -50 "$f" | grep -m1 "^confidence:" | sed 's/^confidence:[[:space:]]*//' || true)
  if [ -z "$doc_title" ]; then
    doc_title=$(head -20 "$f" | grep -m1 "^# " | sed 's/^# //')
  fi
  printf "| [%s](%s) | %s | %s | %s |\n" \
    "${fname%.md}" "$f" "${doc_title:-(untitled)}" "${doc_status:--}" "${doc_confidence:--}"
done

If arg is all

Show counts and status breakdown for each document type:

echo "## Blueprint Documents Summary"
echo ""
for doc_type in adrs prds prps; do
  doc_count=$(ls docs/$doc_type/*.md 2>/dev/null | grep -cv 'README.md' || echo 0)
  echo "### ${doc_type^^}: $doc_count documents"
  if [ "$doc_count" -gt 0 ]; then
    for s in Accepted Proposed Deprecated Superseded draft ready approved; do
      sc=$(grep -ril "^status:.*$s\|^$s$" docs/$doc_type/*.md 2>/dev/null | wc -l | tr -d ' ')
      [ "$sc" -gt 0 ] && echo "- $s: $sc"
    done
  fi
  echo ""
done

Post-Actions

After listing, suggest:

  • For empty directories: "Run /blueprint:derive-adr or /blueprint:derive-prd to generate documents"
  • For stale documents: "Review documents with status 'draft' or 'Proposed'"