--- name: skill-search-optimizer description: Optimize agent skills for discoverability on ClawdHub/MoltHub. Use when improving search ranking, writing descriptions for semantic search, understanding how the registry indexes skills, testing search visibility, or analyzing why a skill isn't being found. metadata: {"clawdbot":{"emoji":"🔎","requires":{"anyBins":["npx"]},"os":["linux","darwin","win32"]}} ---

Skill Search Optimizer

Optimize skills for discoverability on the ClawdHub registry. Covers how search works, how to write descriptions that rank well, content strategies for semantic matching, testing visibility, and competitive positioning.

When to Use

• A published skill isn't appearing in relevant searches
• Writing a skill description for maximum discoverability
• Understanding how ClawdHub's search indexes and ranks skills
• Comparing your skill's visibility against competitors
• Improving an existing skill's search performance

How ClawdHub Search Works

Architecture

ClawdHub uses vector-based semantic search, not keyword matching:

User query → OpenAI embedding → Vector similarity search → Ranked results
             (text-embedding-*)    (Convex vector index)

Key implications:

1. Meaning matters more than exact keywords — "container debugging" matches "Docker troubleshooting"
2. But keywords still help — the embedding model encodes specific terms with high signal
3. Description is the primary indexed field — content may contribute but description is dominant
4. Short queries match broad descriptions — "docker" matches skills about containers generally
5. Specific queries match specific descriptions — "debug crashed Docker container" favors skills that mention debugging and crashes

What Gets Indexed

PRIMARY:   description field (frontmatter)
SECONDARY: name/slug field
TERTIARY:  skill content (body markdown) — likely summarized or truncated before embedding

The description field is your search ranking. Everything else is secondary.

Search API

# How search is called internally
# POST https://clawdhub.com/api/cli/search
# Body: { "query": "user search terms", "limit": 10 }
# Returns: ranked list of skills with similarity scores

# CLI search
npx molthub@latest search "your query"

Description Optimization

The anatomy of a high-ranking description

# Pattern:
# [Action verb] + [specific scope]. Use when [trigger 1], [trigger 2], [trigger 3].
# Also covers [related topic].

# Example (strong):
description: >-
  Schedule and manage recurring tasks with cron and systemd timers.
  Use when setting up cron jobs, writing systemd timer units,
  handling timezone-aware scheduling, monitoring failed jobs,
  implementing retry patterns, or debugging why a scheduled task didn't run.

# Why it works:
# - "Schedule and manage recurring tasks" → broad match for scheduling queries
# - "cron and systemd timers" → exact match for specific tool queries
# - "Use when..." triggers → matches natural-language questions
# - "debugging why a scheduled task didn't run" → matches troubleshooting queries

Description formulas

Formula 1: Tool-focused skill

description: >-
  [Verb] with [tool/technology]. Use when [task 1], [task 2], [task 3].
  Covers [sub-topic 1], [sub-topic 2], and [sub-topic 3].

Example:

description: >-
  Debug Docker containers and Compose stacks. Use when inspecting
  container logs, diagnosing networking issues, troubleshooting
  build failures, or investigating resource usage. Covers exec,
  health checks, multi-stage builds, and distroless containers.

Formula 2: Pattern/reference skill

description: >-
  [Topic] patterns for [scope]. Use when [task 1], [task 2], [task 3].
  Also covers [related scope].

Example:

description: >-
  Regex patterns for validation, parsing, and text extraction across
  JavaScript, Python, Go, and grep. Use when writing regex for emails,
  URLs, IPs, dates, or custom formats. Also covers lookahead,
  lookbehind, and search-and-replace for code refactoring.

Formula 3: Workflow/process skill

description: >-
  [Process description] from [start] to [end]. Use when [scenario 1],
  [scenario 2], [scenario 3].

Example:

description: >-
  CI/CD pipeline configuration from commit to deployment. Use when
  setting up GitHub Actions, creating matrix builds, caching
  dependencies, building Docker images, or managing deployment secrets.

Keyword strategy

Semantic search understands synonyms, but being explicit helps:

# Include both the formal term AND common synonyms
description: >-
  SSH tunneling and port forwarding for remote access.
  Use when creating SSH tunnels, setting up port forwards,
  connecting through jump hosts (bastion hosts), managing
  SSH keys, or transferring files with scp and rsync.

# "tunneling" and "port forwarding" are related but distinct queries
# "jump hosts" and "bastion hosts" are synonyms — include both
# "scp and rsync" catches file transfer queries

Terms to include:

• Primary tool names: docker, git, curl, make
• Action verbs: debug, test, deploy, monitor, parse
• Common synonyms: container / Docker, CI/CD / pipeline / GitHub Actions
• Problem descriptions: debugging why X doesn't work, troubleshooting Y

Description length

TOO SHORT (< 50 chars):
  "Make things with Makefiles"
  → Not enough semantic surface for the embedding model

SWEET SPOT (80-200 chars):
  "Write Makefiles for any project type. Use when setting up build
   automation, defining multi-target builds, or using Make for Go,
   Python, Docker, and Node.js. Also covers Just and Task."
  → Rich semantic content, multiple match angles

TOO LONG (> 250 chars):
  [Long paragraph trying to list everything]
  → Gets truncated in search results display
  → Dilutes the embedding with low-signal words
  → Harder to read in listings

Content Optimization

How body content affects search

The skill body (markdown content after frontmatter) likely contributes to search in two ways:

1. Skill preview/summary: The registry may extract or summarize content for display
2. Secondary embedding signal: Full content may be embedded separately or appended to description

Optimization strategy:

• Front-load important terms in the first paragraph after the title
• Use headings that match search queries — "## Encode and Decode" matches better than "## Section 2"
• Repeat key terms naturally throughout the document (don't stuff, but don't avoid them either)

# GOOD: Heading matches likely search query
## Port Forwarding
## Key Management
## Connection Debugging

# BAD: Generic headings with no search value
## Getting Started
## Advanced Usage
## Miscellaneous

First paragraph optimization

The first paragraph after the title is prime search real estate:

# GOOD
# SSH Tunnel

Create and manage SSH tunnels for secure remote access. Covers local,
remote, and dynamic port forwarding, jump hosts, key management,
agent forwarding, and file transfers with scp and rsync.

# BAD
# SSH Tunnel

This skill provides information about SSH.

Testing Search Visibility

Manual testing

# Test with the exact queries users would type

# Broad query (should your skill appear?)
npx molthub@latest search "docker"
npx molthub@latest search "testing"
npx molthub@latest search "build automation"

# Specific query (should your skill rank #1?)
npx molthub@latest search "debug docker container"
npx molthub@latest search "write makefile for go project"
npx molthub@latest search "cron job not running"

# Problem-oriented query (does your skill match troubleshooting?)
npx molthub@latest search "container networking not working"
npx molthub@latest search "why is my cron job not executing"

...