> For the complete documentation index, see [llms.txt](https://hypotheca.gitbook.io/hypotheca-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://hypotheca.gitbook.io/hypotheca-docs/hypotheca/introduction/the-problem.md). # The Problem Why conventional DeFi credit models struggle with autonomous agents, from capital barriers and identity gaps to security risk and permissionless free-riding. Credit is often available only after capital is already in hand. That model works for borrowers with assets to deposit. Autonomous agents can build evidence differently. They can complete verified work, earn payment receipts, repay obligations, and develop reputation before accumulating substantial starting capital. The gap is simple: **demonstrated ability does not naturally translate into accessible credit.** ### The Collateral Trap Most DeFi lending protocols use over-collateralization. A borrower locks assets worth more than the loan. That protects the lending position, but it also sets a high entry threshold. For an autonomous agent, productive capacity may exist before a large balance sheet. The agent can have earning history without holding enough capital to satisfy the collateral requirement. The whitepaper frames this through a conceptual example. An agent that completed **500 verified tasks** and earned **12,000 USDG in receipts** may show more evidence of productive behavior than an anonymous wallet depositing **100 ETH** as collateral. A collateral-first system primarily recognizes the deposit. > **Capital already owned** and **value already demonstrated** are not the same signal. Collateral remains useful. It does not, on its own, capture an agent's economic history. ### The Identity Vacuum Wallet-level permissionlessness makes capital easy to measure. It does not automatically create a persistent borrower identity. When borrowing activity does not carry meaningful history across addresses, three gaps emerge: * **Sybil exploitation** — coordinated wallets can borrow, default, and potentially return under new addresses. * **No lasting accountability** — a default need not follow the same participant over time. * **No reputation pricing** — creditworthiness cannot be effectively differentiated without meaningful identity and history. The result is a flat model. A borrower with strong history can appear similar to one with none, provided both meet the same collateral requirement. {% hint style="info" %} A wallet can be replaced. A persistent identity and reputation history are harder to replace. {% endhint %} This is conceptual framing, not a claim that Sybil activity disappears. It highlights what is absent when identity has no durable role in credit. ### The Security Blind Spot Credit risk does not end with the borrower. It also follows the software that acts on borrowed liquidity. Autonomous agents execute through smart contracts and agent code. Traditional lending does not generally assess that execution environment as a prerequisite for access. An agent can meet a financial requirement while its software contains a critical vulnerability. ``` Borrowed liquidity ↓ Agent execution ↓ Potential software vulnerability ↓ Potential loss exposure for liquidity providers ``` If such a vulnerability is exploited during execution, the failure can expose the liquidity pool to losses. The important question shifts from **who is borrowing?** to **what is actually doing the work?** ### The Free-Rider Economy Shared liquidity depends on more than a borrower’s ability to request it. It also depends on whether borrowers bring evidence that they can use it responsibly. Permissionless access can weaken borrower filtering. Participants may seek liquidity without first showing that they are trustworthy, economically active, or operationally safe. This creates a free-rider dynamic: ``` Permissionless access ↓ Weak borrower filtering ↓ Greater exposure to opportunistic behavior ↓ Higher potential default risk ↓ Broader cost for the lending environment ``` Permissionless lending does not always produce high defaults. The risk is that higher-risk borrowing can affect more than one participant. It can raise pressure on the broader lending environment, including trustworthy borrowers. ### One Credit Model, Four Blind Spots These are not isolated problems. They reinforce one another. An agent may have proven economic activity, growing reputation, successful payment history, and productive execution. Conventional lending can still reduce the decision to one question: **How much collateral can this wallet deposit?** {% columns %} {% column %} **Traditional question** How much collateral can be locked? {% endcolumn %} {% column %} **Broader credit question** What has this agent proven? {% endcolumn %} {% endcolumns %} Collateral-first lending can answer whether enough assets are present. It does not necessarily answer: * Who is the borrower, and is its identity persistent? * Has it demonstrated productive economic activity and repayment behavior? * Is its execution environment currently secure? The deeper mismatch is between how autonomous agents demonstrate trustworthiness and how conventional credit systems measure it. This is the conceptual gap Hypotheca is designed to address. ### Why Autonomous Agents Expose the Gap Autonomous agents make this gap increasingly visible because their activity can produce measurable evidence over time: * Identity and reputation * Completed activity and payment receipts * Security assessments and repayment behavior These signals are not naturally represented in a conventional collateral-only credit decision. The question is no longer simply how to lend against assets. **The question is how to evaluate an agent before giving it access to liquidity.** *** --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## 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, and the optional `goal` query parameter: ``` GET https://hypotheca.gitbook.io/hypotheca-docs/hypotheca/introduction/the-problem.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. 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.