# Threat Model

This model highlights realistic failure/attack scenarios and the intended mitigations.

#### A. Session agent key compromise

**Scenario:** An attacker gains control of the agent signer.

**Mitigations**

* Recipient pinning prevents arbitrary destination changes
* Allowlists prevent calling arbitrary contracts/tokens
* Rolling budgets cap cumulative loss (reverts beyond cap)

**Operator response**

* Pause agent
* Rotate/revoke session key
* Withdraw/reduce exposure until confident

#### B. Oracle failure or manipulation

**Scenario:** Oracle providers revert, are misconfigured, or return incorrect values.

**Mitigations**

* Fallback provider logic (main → backup)
* Revert when both fail, or the asset is unsupported

**Operator response**

* Pause agent
* Withdraw funds from the pool to the user's Smart Wallet if needed
* Do not widen slippage to force execution during Oracle instability

#### C. External protocol risk (AMM/router)

**Scenario:** Underlying AMM contracts or routers behave unexpectedly or are attacked.

**Mitigations**

* Restrict targets/selectors to known integrations
* Enforce min-out and deviation budgets

**Operator response**

* Pause agent
* Withdraw funds from the pool to the user's Smart Wallet if needed
* Follow incident communications

#### D. Admin key risk/configuration risk

**Scenario:** Admin-controlled parameters are misconfigured or compromised.

**Mitigations**

* Minimize admin surface area
* Operational controls (multisig, timelocks, reviews) — document explicitly

**Operator response**

* If an incident is suspected, pause the agent & contracts
* Verify official communications and on-chain changes


---

# 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.alloc8.xyz/security/security-model/threat-model.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.