prompt-find

λ(keywords) → search_results | keywords := "$@"

search :: Keywords → Results search(K) = { library: get_library_path(), if (not exists(library)): display: "No prompt library found. Save your first prompt with '/meta Refine prompt: <text>'", return: empty, files: glob(library + "*.md"), if (empty(files)): display: "Library is empty. Save your first prompt with '/meta Refine prompt: <text>'", return: empty, matches: ∀file ∈ files: { metadata: parse_frontmatter(file), content: read_file(file), score: calculate_match_score(K, metadata, content) } where score > 0, sorted: sort_desc(matches, m → m.score), display: format_results(sorted, K), if (|sorted| > 0): offer: "Show full prompt? Enter ID or 'q' to quit" } get_library_path :: () → Path get_library_path() = project_root() + "/.meta-cc/prompts/library/" parse_frontmatter :: FilePath → Metadata parse_frontmatter(F) = { # Extract YAML frontmatter (lines 2-12 typically) # Fields: id, title, category, keywords, usage_count, updated } calculate_match_score :: (Keywords, Metadata, Content) → Score calculate_match_score(K, M, C) = { keyword_matches: count_matches(K, M.keywords ∪ extract_text(M.title)), content_matches: count_matches(K, C.original_prompts), category_matches: count_matches(K, M.category), score: (keyword_matches 3) + (content_matches 2) + (category_matches * 1) } format_results :: ([Matches], Keywords) → Display format_results(M, K) = { header: "Found " + |M| + " prompts matching: " + K, separator: "─" * 80, table: ∀match ∈ M: format_row(match), footer: "\nUse '/prompt-show <id>' to view full prompt" } format_row :: Match → String format_row(M) = sprintf( "%-40s %-15s %-8s %s", M.metadata.id, M.metadata.category, "★" * min(5, M.score / 2), # Star rating truncate(M.metadata.title, 50) )

Execute the following steps: • Check library exists: `bash if [ ! -d .meta-cc/prompts/library/ ]; then echo "No prompt library found." echo "Save your first prompt with: /meta Refine prompt: <your-prompt>" exit 0 fi ` • Search prompts: • Read all .md files in library • For each file: • Parse frontmatter (id, title, category, keywords) • Extract original prompts section • Calculate match score based on keyword overlap • Sort by score descending • Display top matches • Display results: • Show table: ID | Category | Score | Title • Provide option to view full prompt with /prompt-show <id> • Example output: ` Found 3 prompts matching: phase execute plan ──────────────────────────────────────────────────────────────────────────────── ID Category Score Title ──────────────────────────────────────────────────────────────────────────────── phase-execution-systematic-plan-001 phase-exec ★★★★★ Systematic Phase Planning and Execution debug-error-analysis-001 debug ★★ Error Analysis and Debugging commit-workflow-001 git ★ Complete Git Commit Workflow Use '/prompt-show <id>' to view full prompt `

Search keywords: $@ If empty, display usage: ` Usage: /prompt-find <keywords> Example: /prompt-find phase plan execute `