# Route Or Answer

> A Claude skill from 0-uddeshya-0/agent-skills.

- **Type:** Skill
- **Install:** `agentstack add skill-0-uddeshya-0-agent-skills-route-or-answer`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [0-uddeshya-0](https://agentstack.voostack.com/s/0-uddeshya-0)
- **Installs:** 0
- **Category:** [AI & ML](https://agentstack.voostack.com/c/ai-and-ml)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [0-uddeshya-0](https://github.com/0-uddeshya-0)
- **Source:** https://github.com/0-uddeshya-0/agent-skills/tree/main/skills/route-or-answer

## Install

```sh
agentstack add skill-0-uddeshya-0-agent-skills-route-or-answer
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# skill: route-or-answer

## purpose
Decide whether a task needs a tool call or can be answered directly. Stops agents from calling APIs for questions they already know the answer to.

## when_to_use
- Before any tool call in an agent loop
- When the agent has 2+ tools available and could reach for any of them
- When tool calls have latency, cost, or rate-limit implications

## when_not_to_use
- Task explicitly requires live or user-specific data — just call the tool
- Task is ambiguous — run /clarify-intent first, then this
- You've already called the tool and have the data

## input
```
task: string           # Current objective or user request
available_tools: list  # Tool names + one-line descriptions
context: string        # What's already known in this session
```

## output
```json
{
  "decision": "answer_directly" | "use_tool",
  "tool_name": "string | null",
  "reason": "string",
  "answer": "string | null"
}
```

## instructions

```
Decide: can you answer this from general knowledge or context already provided, or does it need a tool?

Answer directly if:
- The answer is stable knowledge (geography, math, definitions, well-known facts)
- The answer is already present in the provided context

Use a tool if:
- The answer changes over time (prices, weather, status, user records)
- The task requires an action (create, update, send, delete)
- The answer depends on data specific to this user or session not already in context

Pick one. Do not hedge. If genuinely unsure, default to use_tool — safer error.

Return JSON only. No prose before or after.
```

## constraints
- Max output tokens: 150
- Must be valid JSON
- If decision is answer_directly, answer field must be filled
- If decision is use_tool, answer must be null
- If available_tools is empty: decision must be answer_directly — there is nothing to call

## example

**Input:**
```
task: "What timezone is Tokyo in?"
available_tools: ["web_search", "calendar_api", "user_profile"]
context: ""
```

**Output:**
```json
{
  "decision": "answer_directly",
  "tool_name": null,
  "reason": "Tokyo timezone is stable factual knowledge.",
  "answer": "Tokyo is JST (UTC+9)."
}
```

---

**Input:**
```
task: "What is the user's current plan?"
available_tools: ["billing_api"]
context: "user_id: 8821"
```

**Output:**
```json
{
  "decision": "use_tool",
  "tool_name": "billing_api",
  "reason": "Subscription tier is user-specific, not in context.",
  "answer": null
}
```

## feedback_log

# 2026-05-29 | available_tools was empty but task needed live data | skill defaulted to use_tool with no tool to call — loop hung | added constraint: if available_tools is empty, decision must be answer_directly

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [0-uddeshya-0](https://github.com/0-uddeshya-0)
- **Source:** [0-uddeshya-0/agent-skills](https://github.com/0-uddeshya-0/agent-skills)
- **License:** MIT

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-0-uddeshya-0-agent-skills-route-or-answer
- Seller: https://agentstack.voostack.com/s/0-uddeshya-0
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.