# Core Components

The Pilot Agent Control Layer is built on five key components.\
Together, these modules enable agents to understand intent, route actions securely, and execute across chains in a transparent and scalable way.

***

### 1. Intent Parsing

The starting point of every action.

* Translates natural language input into structured commands.
* Identifies *what the user wants* (swap, transfer, deposit, etc.) and the *parameters* (amount, asset, destination).
* Uses LLMs and rule-based logic to ensure commands are precise enough to execute.

**Example:**\
User: *“Swap 100 USDC to ETH and send it to Alice.”*\
→ Parsed as `{ action: swap, amount: 100, token: USDC, output: ETH, transferTo: Alice }`.

***

### 2. Action Router

Responsible for connecting intent to the right onchain service.

* Chooses the best protocol, chain, or integration based on availability and user context.
* Optimizes for liquidity, fees, and speed.
* Modular → easily extended with new adapters for protocols and chains.

**Example:**\
Swap request can be routed via:

* Cetus (Sui)
* 0x API (EVM)
* Jupiter (Solana)

***

### 3. Execution Layer

The transaction engine that turns routed actions into onchain results.

* Builds transactions in the required format (PTBs on Sui, bundled calls on EVM, etc.).
* Handles gasless execution where supported.
* Confirms and tracks the state of execution.

**Key point:** Users don’t deal with raw transactions — the execution layer abstracts all chain-specific complexity.

***

### 4. Wallet Abstraction (AA + Session Keys)

Simplifies wallet management for both users and agents.

* **Account Abstraction (AA):** Enables smart wallets with programmable permissions.
* **Session Keys:** Temporary, scoped keys that allow agents to act without asking the user to sign every step.
* Supports social login (zkLogin, Privy, etc.) for onboarding without seed phrases.

**Benefit:** Agents can operate safely within defined limits (gas caps, token allowances) while maintaining user sovereignty.

***

### 5. Cross-Chain Router

Makes the system chain-agnostic by default.

* Handles asset transfers and liquidity routing between chains.
* Unifies balances so users see *one portfolio*, not fragmented assets.
* Designed to plug into bridges and messaging layers for full interoperability.

**Outcome:** Agents can execute multi-chain strategies without forcing users to switch networks or approve multiple times.

***

### 6. Guardrails & Observability

The trust layer of Pilot.

* Every action is logged and auditable.
* Policies define what agents *can and cannot do* (limits, frequency, token allowlists/blacklists).
* Observability dashboards provide transparency for users, developers, and protocols.

**Example:**\
If an agent swaps 1000 USDC, the log will show:

* Source (wallet ID)
* Action (swap)
* Route (protocol used)
* Result (tokens received, fees paid)


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.pilotagent.org/ai-agent-control-layer/core-components.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.