An agent that cannot remember what it approved feels forgetful. An agent that stores everything without control feels invasive. Nerva’s planned local SQLite store can provide a middle path: keep useful task history on the device, make retention understandable, and avoid turning memory into an unlimited dump of private data.

What belongs in local storage

A first schema could track:

Record Example fields
Task title, state, created time, workspace
Step proposal, status, tool name, result
Approval action, scope, decision, expiry
Context selected paths, redactions, destination
Verification command, exit status, evidence

The database does not need to store every full file or model response. It can store references, summaries, hashes, and user-selected notes. That keeps the history useful without quietly duplicating a repository.

Local does not mean permanent

Nerva should let users delete task history, set retention rules, and see where local data lives. A local database can still contain sensitive filenames, error messages, and personal notes. The product should not imply that “on device” means “risk free.”

SQLite also gives the task manager reliable state transitions. A task can move from waiting for approval to running, failed, or verified without relying on a fragile in-memory object. This matters after a crash or restart: Nerva should know whether a command actually ran before offering to repeat it.

For a developer reviewing work on CryptoEx or SentinelFi, a compact local record of selected files, decisions, and test results could be more valuable than a long chat transcript. The storage layer should preserve that evidence while keeping control in the user’s hands.

The useful design question

The useful question here is not whether an AI agent can produce an impressive demo. It is whether the workflow stays understandable when the task is incomplete, the repository is messy, or the model is wrong. For this part of Nerva, the focus is making local history useful without turning it into an uncontrolled memory store. That means each state should have a visible owner, a small input, and a result that another part of the system can inspect.

A concrete scenario

Consider saving task steps, approvals, and verification results while excluding secrets. A good Nerva run would first state the scope, identify the files or operation it needs, and show the next proposed step. It would keep the original request separate from model-generated instructions. If an action needs approval, the user would see what will happen before it happens. Afterward, the task history would retain the command, result, and verification status instead of reducing everything to a confident paragraph.

This approach also makes a failure useful. A denied permission, unavailable provider, missing package, or changed file is information about the task state. Nerva should preserve it and explain the smallest recovery option. It should not silently broaden the request, retry a side effect, or claim completion because a model supplied plausible-looking text.

What I would test first

For an early implementation, I would test three paths: a normal successful run, a denied action, and an interrupted run resumed after the workspace changes. The key test is how long each record should live and how it can be deleted. In each path, the interface should make clear what Nerva knows, what it proposed, what actually ran, and what remains unverified.

Useful acceptance checks include:

  • the user can inspect the selected context before model execution;
  • sensitive paths and credentials are excluded or redacted;
  • every side effect has a narrow permission decision;
  • failures remain attached to the step that caused them;
  • a final answer links to evidence instead of implying that unrun work succeeded.

Nerva is still an upcoming project, so these are design and implementation goals rather than claims that a finished product already provides them. That distinction matters: a trustworthy agent should earn confidence through observable behavior, not through a feature list.

The boundary that matters in production

SQLite is a good fit for local history because the data is relational: one task owns many steps, each step may have an approval, and each verification result belongs to a concrete action. The database should still have migrations, bounded retention, and a clear reset path. Local does not mean immortal; users need to understand and control what is remembered.

For the first vertical slice, keep the boundary small enough to explain in a code review. Define the input, the decision, the side effect, and the evidence for completion. Add a test for the denied path as well as the successful path. This is where a local-first design earns trust: the user can inspect the context and history on the device, see when a provider is involved, and stop before an action crosses the permission boundary. The design should remain honest about its limits. Nerva is planned software, not a claim that every workflow can be made safe automatically. A clear “cannot verify” state is more useful than a polished but unsupported success message.

A review of this design should use a small fixture repository and record both allowed and rejected outcomes. That keeps the discussion concrete: reviewers can compare the requested scope with the selected context, action, and evidence. It also prevents Nerva from being judged only by how persuasive its response sounds.

The design is informed by the boundaries I have had to think through in CryptoEx and SentinelFi: backend work becomes easier to review when context, permissions, side effects, and verification are explicit.