Practical notes: Building AI Agents in Rust — part 2

The following notes reconstruct a practical path around “Building AI Agents in Rust — part 2”. Emphasis stays on contracts, checks, and drop-in code placeholders rather than motivational framing. When working through the Overview stage, write down the contract first: required inputs, success signal, and what happens on partial failure. That checklist keeps later code changes honest. Treat this stage as a contract between inputs and validated outputs. Name the artifacts, define success checks, and refuse silent partial completion.

Why structure matters here

The Why structure matters here stage works best when treated as a measurable surface. Capture one golden transcript, one failure case, and the rollback note before expanding scope. Record timings and token or query cost next to functional results. Cost visibility early prevents surprise bills when the path moves from demo to shared environments. Keep graph state flat and typed. Nested blobs hide which node wrote which field and break resume after interrupts.

The four layers

The The four layers stage works best when treated as a measurable surface. Capture one golden transcript, one failure case, and the rollback note before expanding scope. Keep configuration outside application code. Environment files, secret stores, and feature flags belong in one place operators can audit without reading the whole graph. Keep graph state flat and typed. Nested blobs hide which node wrote which field and break resume after interrupts.

A typed builder

The A typed builder stage works best when treated as a measurable surface. Capture one golden transcript, one failure case, and the rollback note before expanding scope. Document the happy path and the recovery path together. Retries, human gates, and dead-letter handling are part of the product, not later polish. Keep graph state flat and typed. Nested blobs hide which node wrote which field and break resume after interrupts. The A typed builder stage works best when treated as a measurable surface. Capture one golden transcript, one failure case, and the rollback note before expanding scope. Treat this stage as a contract between inputs and validated outputs. Name the artifacts, define success checks, and refuse silent partial completion.

let prompt = SystemPromptBuilder::new()
    .identity("You are Eugene, a careful research assistant who answers \
               questions about a Rust project. You prefer reading the source \
               over guessing.")
    .instruction("Use `list_files` to discover what is in the project before reading.")
    .instruction("Use `read_file` to inspect a specific file. Do not call it on \
                  paths you have not seen listed.")
    .instruction("If a tool returns an error, do not retry the same call.")
    .output_constraints("Answer in plain prose. Cite the file you read in parentheses, \
                         for example: (src/main.rs).")
    .example("What edition does Cargo.toml use?",
             "I'll check Cargo.toml directly. (Cargo.toml) The project uses Rust edition 2024.")
    .context(format!("<env>\ntoday: {today}\nproject_root: {sandbox}\n</env>"))
    .build();

## Identity

You are Eugene, a careful research assistant ...

## Instructions

- Use `list_files` to discover what is in the project before reading.
- Use `read_file` to inspect a specific file. Do not call it on paths ...
- If a tool returns an error, do not retry the same call.

## Output

Answer in plain prose. Cite the file you read in parentheses ...

## Examples

Example 1:
User: What edition does Cargo.toml use?
Assistant: I'll check Cargo.toml directly. (Cargo.toml) ...

## Context

<env>
today: 2026-05-22
project_root: /Users/me/code/eugene
</env>

Identity is voice, not truth

For the Identity is voice not stage, define the inputs, the owner of the step, and the exit criteria before changing code. Operators should be able to re-run the step from a known checkpoint without guessing hidden state. Record timings and token or query cost next to functional results. Cost visibility early prevents surprise bills when the path moves from demo to shared environments. Put human approval on edges that spend money or change production data. Compile-time wiring does not equal business completeness.

Instructions are a list, not an essay

For the Instructions are a list stage, define the inputs, the owner of the step, and the exit criteria before changing code. Operators should be able to re-run the step from a known checkpoint without guessing hidden state. Keep configuration outside application code. Environment files, secret stores, and feature flags belong in one place operators can audit without reading the whole graph. Put human approval on edges that spend money or change production data. Compile-time wiring does not equal business completeness.

Output constraints decouple format from behaviour

For the Output constraints decouple format stage, define the inputs, the owner of the step, and the exit criteria before changing code. Operators should be able to re-run the step from a known checkpoint without guessing hidden state. Document the happy path and the recovery path together. Retries, human gates, and dead-letter handling are part of the product, not later polish. Put human approval on edges that spend money or change production data. Compile-time wiring does not equal business completeness. For the Output constraints decouple format stage, define the inputs, the owner of the step, and the exit criteria before changing code. Operators should be able to re-run the step from a known checkpoint without guessing hidden state. Treat this stage as a contract between inputs and validated outputs. Name the artifacts, define success checks, and refuse silent partial completion.

Few-shot examples demonstrate, don’t explain

When working through the Few-shot examples demonstrate don stage, write down the contract first: required inputs, success signal, and what happens on partial failure. That checklist keeps later code changes honest. Record timings and token or query cost next to functional results. Cost visibility early prevents surprise bills when the path moves from demo to shared environments. Checkpoint after expensive steps. Resume should not re-bill the same LLM call when an operator retries a later node.

.example(
    "What edition does Cargo.toml use?",
    "I'll check Cargo.toml directly. (Cargo.toml) The project uses Rust edition 2024.",
)

Context is fresh on every request

When working through the Context is fresh on stage, write down the contract first: required inputs, success signal, and what happens on partial failure. That checklist keeps later code changes honest. Keep configuration outside application code. Environment files, secret stores, and feature flags belong in one place operators can audit without reading the whole graph. Checkpoint after expensive steps. Resume should not re-bill the same LLM call when an operator retries a later node.

let today = OffsetDateTime::now_utc()
    .date()
    .format(&Iso8601::DATE)
    .unwrap_or_else(|_| "unknown".into());
let sandbox = sandbox_root()?.display().to_string();

let prompt = system_prompt(&today, &sandbox);

The cache boundary

When working through the The cache boundary stage, write down the contract first: required inputs, success signal, and what happens on partial failure. That checklist keeps later code changes honest. Document the happy path and the recovery path together. Retries, human gates, and dead-letter handling are part of the product, not later polish. Checkpoint after expensive steps. Resume should not re-bill the same LLM call when an operator retries a later node. When working through the The cache boundary stage, write down the contract first: required inputs, success signal, and what happens on partial failure. That checklist keeps later code changes honest. Treat this stage as a contract between inputs and validated outputs. Name the artifacts, define success checks, and refuse silent partial completion.

let blocks = prompt.into_system_blocks();
// blocks[0] = { type: "text", text: <static prefix>, cache_control: {type: "ephemeral"} }
// blocks[1] = { type: "text", text: <dynamic suffix> }   // no cache_control

[turn 0] in=4 cache_read=0 cache_create=1247 out=89
[turn 1] in=3 cache_read=1247 cache_create=0 out=42

Section memoization: compute once, reuse

The Section memoization compute once stage works best when treated as a measurable surface. Capture one golden transcript, one failure case, and the rollback note before expanding scope. Record timings and token or query cost next to functional results. Cost visibility early prevents surprise bills when the path moves from demo to shared environments. Keep graph state flat and typed. Nested blobs hide which node wrote which field and break resume after interrupts.

Versioning and regression testing

The Versioning and regression testing stage works best when treated as a measurable surface. Capture one golden transcript, one failure case, and the rollback note before expanding scope. Keep configuration outside application code. Environment files, secret stores, and feature flags belong in one place operators can audit without reading the whole graph. Keep graph state flat and typed. Nested blobs hide which node wrote which field and break resume after interrupts.

const EXPECTED_PROMPT_FINGERPRINT: u64 = 0; // set after first successful run

if EXPECTED_PROMPT_FINGERPRINT != 0
    && prompt.fingerprint() != EXPECTED_PROMPT_FINGERPRINT
{
    eprintln!(
        "warning: system prompt fingerprint drifted (was {EXPECTED_PROMPT_FINGERPRINT}, \
         is {}). Update the constant if the change was intentional.",
        prompt.fingerprint()
    );
}

Eugene v0.2 in practice

The Eugene v0 2 in stage works best when treated as a measurable surface. Capture one golden transcript, one failure case, and the rollback note before expanding scope. Document the happy path and the recovery path together. Retries, human gates, and dead-letter handling are part of the product, not later polish. Keep graph state flat and typed. Nested blobs hide which node wrote which field and break resume after interrupts. The Eugene v0 2 in stage works best when treated as a measurable surface. Capture one golden transcript, one failure case, and the rollback note before expanding scope. Treat this stage as a contract between inputs and validated outputs. Name the artifacts, define success checks, and refuse silent partial completion.

let prompt = system_prompt(&today, &sandbox);
let system_blocks = prompt.into_system_blocks();

let response = send(&http, &api_key, &system_blocks, &tools, &messages).await?;

What this reveals

For the What this reveals stage, define the inputs, the owner of the step, and the exit criteria before changing code. Operators should be able to re-run the step from a known checkpoint without guessing hidden state. Record timings and token or query cost next to functional results. Cost visibility early prevents surprise bills when the path moves from demo to shared environments. Put human approval on edges that spend money or change production data. Compile-time wiring does not equal business completeness.

What comes next

For the What comes next stage, define the inputs, the owner of the step, and the exit criteria before changing code. Operators should be able to re-run the step from a known checkpoint without guessing hidden state. Keep configuration outside application code. Environment files, secret stores, and feature flags belong in one place operators can audit without reading the whole graph. Put human approval on edges that spend money or change production data. Compile-time wiring does not equal business completeness.

The workspace

For the The workspace stage, define the inputs, the owner of the step, and the exit criteria before changing code. Operators should be able to re-run the step from a known checkpoint without guessing hidden state. Document the happy path and the recovery path together. Retries, human gates, and dead-letter handling are part of the product, not later polish. Put human approval on edges that spend money or change production data. Compile-time wiring does not equal business completeness. For the The workspace stage, define the inputs, the owner of the step, and the exit criteria before changing code. Operators should be able to re-run the step from a known checkpoint without guessing hidden state. Treat this stage as a contract between inputs and validated outputs. Name the artifacts, define success checks, and refuse silent partial completion.

Related topics

When working through the Related topics stage, write down the contract first: required inputs, success signal, and what happens on partial failure. That checklist keeps later code changes honest. Record timings and token or query cost next to functional results. Cost visibility early prevents surprise bills when the path moves from demo to shared environments. Checkpoint after expensive steps. Resume should not re-bill the same LLM call when an operator retries a later node.

The code

When working through the The code stage, write down the contract first: required inputs, success signal, and what happens on partial failure. That checklist keeps later code changes honest. Keep configuration outside application code. Environment files, secret stores, and feature flags belong in one place operators can audit without reading the whole graph. Checkpoint after expensive steps. Resume should not re-bill the same LLM call when an operator retries a later node.

Want more like this?

When working through the Want more like this stage, write down the contract first: required inputs, success signal, and what happens on partial failure. That checklist keeps later code changes honest. Document the happy path and the recovery path together. Retries, human gates, and dead-letter handling are part of the product, not later polish. Checkpoint after expensive steps. Resume should not re-bill the same LLM call when an operator retries a later node.

Operational checklist

For the Operational checklist stage, define the inputs, the owner of the step, and the exit criteria before changing code. Operators should be able to re-run the step from a known checkpoint without guessing hidden state.

Prefer small, testable units over sprawling scripts. When a step fails, the failure should point at a single responsibility rather than a tangled pipeline.

Put human approval on edges that spend money or change production data. Compile-time wiring does not equal business completeness.

Write a short runbook: how to rotate keys, how to drain the queue, how to roll back the last ingest.

Treat this stage as a contract between inputs and validated outputs. Name the artifacts, define success checks, and refuse silent partial completion.

Put human approval on edges that spend money or change production data. Compile-time wiring does not equal business completeness.

Before promoting the stack, freeze versions, capture a golden transcript for the critical path, and confirm rollback steps. Shared environments need rate limits, tenancy checks, and a clear owner for secret rotation. Prefer boring reliability over clever one-off demos.

Batch note for bfb02d0288c5: keep provider keys out of the repo, set a per-session token ceiling, and store transcripts next to the eval fixtures so later model swaps stay comparable.

For the hardening note 0 stage, define the inputs, the owner of the step, and the exit criteria before changing code. Operators should be able to re-run the step from a known checkpoint without guessing hidden state. Document the happy path and the recovery path together. Retries, human gates, and dead-letter handling are part of the product, not later polish.

Hardening detail 0/858: measure wall time, error class, and token spend for this note, then decide whether to keep the change based on a fixed question set rather than anecdote.

When working through the hardening note 1 stage, write down the contract first: required inputs, success signal, and what happens on partial failure. That checklist keeps later code changes honest. Treat this stage as a contract between inputs and validated outputs. Name the artifacts, define success checks, and refuse silent partial completion.

Hardening detail 1/858: measure wall time, error class, and token spend for this note, then decide whether to keep the change based on a fixed question set rather than anecdote.