router-builder
Build intent routers for agent task distribution
What this skill does
# Router Builder Skill
> Build the Semantic Router for intelligent resource selection.
## Overview
The router provides fast, embedding-based routing to:
- Slash commands
- Sub-agents
- Skills
- Workflows
Uses the same embedding model as RAG for consistency.
## Prerequisites
```bash
pip install semantic-router sentence-transformers pyyaml
```
## Build Steps
### Step 1: Create Router Core
**File: `routing/router.py`**
```python
#!/usr/bin/env python3
"""
Hierarchical Semantic Router
Two-tier routing:
- Tier 1: Category (command | agent | skill | workflow)
- Tier 2: Specific resource within category
"""
import os
import yaml
from pathlib import Path
from typing import Optional, Dict, List, Tuple
from dataclasses import dataclass
from enum import Enum
from semantic_router import Route
from semantic_router.layer import RouteLayer
from semantic_router.encoders import HuggingFaceEncoder
# Configuration
ROUTES_PATH = Path(os.getenv("ROUTES_PATH", "./routing/routes"))
CONFIDENCE_THRESHOLD = float(os.getenv("ROUTING_CONFIDENCE", "0.5"))
class Category(Enum):
COMMAND = "command"
AGENT = "agent"
SKILL = "skill"
WORKFLOW = "workflow"
GENERAL = "general"
@dataclass
class RoutingResult:
"""Result of a routing decision."""
category: Category
resource: Optional[str]
confidence: float
metadata: Dict
fallback: bool = False
class Router:
"""Hierarchical semantic router."""
def __init__(self):
# Use same model as RAG
self.encoder = HuggingFaceEncoder(name="all-MiniLM-L6-v2")
self.category_layer = self._build_category_layer()
self.domain_layers: Dict[Category, RouteLayer] = {}
self._build_domain_layers()
def _load_routes(self, path: Path) -> List[Route]:
"""Load routes from YAML."""
if not path.exists():
return []
with open(path) as f:
data = yaml.safe_load(f) or {}
return [
Route(
name=r["name"],
utterances=r["utterances"],
metadata=r.get("metadata", {})
)
for r in data.get("routes", [])
]
def _build_category_layer(self) -> RouteLayer:
"""Build tier-1 category router."""
routes = self._load_routes(ROUTES_PATH / "categories.yaml")
if not routes:
# Default routes
routes = [
Route(name="command", utterances=[
"/research", "/code-review", "/daily-standup",
"run the command", "execute command", "use slash command"
]),
Route(name="agent", utterances=[
"ask the researcher", "have the coder", "use the analyst",
"delegate to agent", "agent should handle"
]),
Route(name="skill", utterances=[
"use the skill", "apply skill", "leverage skill for"
]),
Route(name="workflow", utterances=[
"run workflow", "start pipeline", "execute automation"
]),
]
return RouteLayer(encoder=self.encoder, routes=routes)
def _build_domain_layers(self):
"""Build tier-2 domain routers."""
mappings = {
Category.COMMAND: "commands.yaml",
Category.AGENT: "agents.yaml",
Category.SKILL: "skills.yaml",
Category.WORKFLOW: "workflows.yaml",
}
for category, filename in mappings.items():
routes = self._load_routes(ROUTES_PATH / filename)
if routes:
self.domain_layers[category] = RouteLayer(
encoder=self.encoder,
routes=routes
)
def route(self, query: str) -> RoutingResult:
"""Route a query through both tiers."""
# Tier 1: Category
cat_result = self.category_layer(query)
if cat_result.name is None:
return RoutingResult(
category=Category.GENERAL,
resource=None,
confidence=0.0,
metadata={"reason": "no_category_match"},
fallback=True
)
category = Category(cat_result.name)
# Tier 2: Specific resource
if category in self.domain_layers:
domain_result = self.domain_layers[category](query)
if domain_result.name is not None:
return RoutingResult(
category=category,
resource=domain_result.name,
confidence=0.8, # TODO: extract actual score
metadata=getattr(domain_result, 'metadata', {}) or {}
)
# Category matched but no specific resource
return RoutingResult(
category=category,
resource=None,
confidence=0.5,
metadata={"reason": "no_resource_match"},
fallback=True
)
def add_route(self, category: Category, name: str, utterances: List[str], metadata: Dict = None):
"""Dynamically add a route."""
route = Route(name=name, utterances=utterances, metadata=metadata or {})
if category not in self.domain_layers:
self.domain_layers[category] = RouteLayer(
encoder=self.encoder,
routes=[route]
)
else:
# Rebuild with new route
existing = list(self.domain_layers[category].routes)
existing.append(route)
self.domain_layers[category] = RouteLayer(
encoder=self.encoder,
routes=existing
)
# Singleton
_router: Optional[Router] = None
def get_router() -> Router:
global _router
if _router is None:
_router = Router()
return _router
def route(query: str) -> RoutingResult:
return get_router().route(query)
```
### Step 2: Create Route Definitions
**File: `routing/routes/categories.yaml`**
```yaml
routes:
- name: command
utterances:
- "/research"
- "/code-review"
- "/daily-standup"
- "/summarize"
- "run the research command"
- "execute code review"
- "use slash command"
- "run command for"
- name: agent
utterances:
- "ask the researcher"
- "have the coder implement"
- "let the writer draft"
- "use the analyst"
- "delegate to agent"
- "which agent should"
- "agent help with"
- name: skill
utterances:
- "use web research skill"
- "apply document generation"
- "use the skill for"
- "leverage skill"
- name: workflow
utterances:
- "run the workflow"
- "start the pipeline"
- "execute automation"
- "trigger workflow"
```
**File: `routing/routes/commands.yaml`**
```yaml
routes:
- name: research
utterances:
- "research this"
- "find information about"
- "look up"
- "investigate"
- "deep dive into"
- "gather info on"
metadata:
file: ".claude/commands/research.md"
- name: code-review
utterances:
- "review code"
- "check for bugs"
- "analyze code"
- "security review"
- "code quality"
metadata:
file: ".claude/commands/code-review.md"
- name: daily-standup
utterances:
- "daily standup"
- "standup report"
- "what did I work on"
- "yesterday today blockers"
metadata:
file: ".claude/commands/daily-standup.md"
```
**File: `routing/routes/agents.yaml`**
```yaml
routes:
- name: researcher
utterances:
- "research topic"
- "find information"
- "look up facts"
- "gather data"
- "fact check"
metadata:
path: "agents/sub-Related in AI Agents
skill-development
IncludedComprehensive meta-skill for creating, managing, validating, auditing, and distributing Claude Code skills and slash commands (unified in v2.1.3+). Provides skill templates, creation workflows, validation patterns, audit checklists, naming conventions, YAML frontmatter guidance, progressive disclosure examples, and best practices lookup. Use when creating new skills, validating existing skills, auditing skill quality, understanding skill architecture, needing skill templates, learning about YAML frontmatter requirements, progressive disclosure patterns, tool restrictions (allowed-tools), skill composition, skill naming conventions, troubleshooting skill activation issues, creating custom slash commands, configuring command frontmatter, using command arguments ($ARGUMENTS, $1, $2), bash execution in commands, file references in commands, command namespacing, plugin commands, MCP slash commands, Skill tool configuration, or deciding between skills vs slash commands. Delegates to docs-management skill for official documentation.
reprompter
IncludedTransform messy prompts into well-structured, effective prompts — single or multi-agent. Use when: "reprompt", "reprompt this", "clean up this prompt", "structure my prompt", rough text needing XML tags and best practices, "reprompter teams", "repromptception", "run with quality", "smart run", "smart agents", multi-agent tasks, audits, parallel work, anything going to agent teams. Don't use when: simple Q&A, pure chat, immediate execution-only tasks. See "Don't Use When" section for details. Outputs: Structured XML/Markdown prompt, quality score (before/after), optional team brief + per-agent sub-prompts, agent team output files. Success criteria: Single mode quality score ≥ 7/10; Repromptception per-agent prompt quality score 8+/10; all required sections present, actionable and specific.
adaptive-compaction
IncludedAdaptive add-on policy and recovery layer that decides WHEN to compact, prune, snapshot, or fork -- replacing fixed-percent auto-compaction across Claude Code, Codex, and MCP-capable hosts. Trigger on auto-compact timing or damage: "when should I compact", "is it safe to compact now or start a fresh session", "auto-compact fires too early/mid-task", "switching to an unrelated task but the window still has space", "context rot", "answers get worse the longer the session runs", "the agent forgot the plan or my decisions after it summarized", "add a layer on top that manages context without changing the agent", raising autoCompactWindow to give the policy room, or installing/tuning a cross-tool compaction policy or PreCompact hook -- even when "compaction" is never said but the problem is context-window pressure or post-summarization memory loss. Do NOT use to summarize a conversation, build RAG, write a summarization prompt (decides WHEN not HOW), or answer max-context-length trivia.
agent-skill-creator
IncludedCreate cross-platform agent skills from workflow descriptions. Activates when users ask to create an agent, automate a repetitive workflow, create a custom skill, or need advanced agent creation. Triggers on phrases like create agent for, automate workflow, create skill for, every day I have to, daily I need to, turn process into agent, need to automate, create a cross-platform skill, validate this skill, export this skill, migrate this skill. Supports single skills, multi-agent suites, transcript processing, template-based creation, interactive configuration, cross-platform export, and spec validation.
llm-wiki
IncludedUse when building or maintaining a persistent personal knowledge base (second brain) in Obsidian where an LLM incrementally ingests sources, updates entity/concept pages, maintains cross-references, and keeps a synthesis current. Triggers include "second brain", "Obsidian wiki", "personal knowledge management", "ingest this paper/article/book", "build a research wiki", "compound knowledge", "Memex", or whenever the user wants knowledge to accumulate across sessions instead of being re-derived by RAG on every query.
skill-master
IncludedAgent Skills authoring, evaluation, and optimization. Create, edit, validate, benchmark, and improve skills following the agentskills.io specification. Use when designing SKILL.md files, structuring skill folders (references, scripts, assets), ingesting external documentation into skills, running trigger evals, benchmarking skill quality, optimizing descriptions, or performing blind A/B comparisons. Keywords: agentskills.io, SKILL.md, skill authoring, eval, benchmark, trigger optimization.