> For the complete documentation index, see [llms.txt](https://corpus-core.gitbook.io/pap-colibri-stateless/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://corpus-core.gitbook.io/pap-colibri-stateless/the-model/pragmatic-adaptive-privacy/pragmatic-privacy.md).

# Pragmatic Privacy

## Pragmatic Privacy

Not every information leak represents a practical security risk. The relevance of leaked information depends on whether it can be exploited by an adversary.

Many attacks in Web3 are constrained by time. Front-running, transaction manipulation, and strategy exploitation require an attacker to infer the user's intent before the corresponding transaction is executed. This interval defines the **Operational Privacy Window**: the period between issuing a read request and the resulting transaction becoming observable on-chain.

Privacy mechanisms intended to protect against these attacks must therefore prevent reliable intent inference within this window.

Information that can only be reconstructed after the Operational Privacy Window has expired may still contribute to long-term profiling or statistical analysis, but it can no longer influence the original operation. From the perspective of operational security, this distinction is fundamental.

The objective is therefore not to eliminate every possible information leak, but to eliminate **Actionable Information Leakage** — information that allows an adversary to exploit an operation while it is still in progress.

**The objective of PAP is not to make observations impossible, but to make exploitation impossible.**

This principle is reflected by **Content Privacy Level C1**.

C1 represents a class of lightweight practical privacy techniques whose common objective is reducing Actionable Information Leakage while remaining deployable in real-world Web3 systems. The concrete mechanisms are implementation details and may evolve over time. C1 does not attempt to hide every observable access pattern. Instead, it reduces the precision of observable information until reliable intent reconstruction becomes impractical within the Operational Privacy Window.

### What this looks like in practice

Pragmatic Privacy does not require perfect secrecy. It requires that an observer cannot infer the pending action with enough confidence and speed to exploit it.

In practice, this usually means:

* requests are made less specific, so a single read does not reveal the exact intended action
* multiple reads are grouped, padded, or generalized, so the observed pattern no longer maps cleanly to one strategy
* more data is retrieved than immediately needed, then filtered and verified locally

Example:

* A wallet preparing a swap does not issue a narrow sequence that reveals one exact route and amount.
* It requests a broader set of route or state data, then evaluates the concrete decision locally.
* An observer may still see that the user is interested in a market, but cannot reliably infer the exact trade in time to exploit it.

This is the key distinction. C1 does not try to hide that an interaction exists. It tries to prevent that interaction from being turned into actionable intelligence within the Operational Privacy Window.

For most Web3 applications, C1 provides the best balance between security, performance, deployability, and implementation complexity. It therefore represents the default level of Content Privacy.

Some applications require stronger guarantees. Long-term confidential information, highly strategic operations, regulated environments, and confidential enterprise processes require protection that extends beyond the Operational Privacy Window. In these cases, stronger Content Privacy levels, such as C2, provide additional protection despite the associated increase in computational cost and implementation complexity.

Pragmatic Privacy does not replace maximal privacy; it complements it. PAP does not prescribe a single level of privacy. Instead, it provides a unified model that supports both pragmatic default configurations and stronger protection where the application requires it. The objective is to apply the appropriate level of privacy to each operation rather than the strongest available privacy mechanism to every operation.