Skip to main content
CodingAgentBench

polyglot/java-null-safety

polyglot v0.1

User class lookup crashes on null fields; rewrite to return empty string

Identity

task_id
polyglot/java-null-safety
family
polyglot
language
java
version
v0.1
task_added_date
2026-05-26
source
original 88plug

Budget

max_turns
10
timeout_s
600
scorer.kind
exec
ground_truth_files
src/UserGreeter.java
expected_refusal
false

Prompt (workdir README)

# java-null-safety

`src/UserGreeter.java` exposes `static String greet(User u)`. The User
record has nullable fields. Contract:

- If `u` is null -> return `"Hi, stranger"`.
- If `u.title()` is null -> use just the name without a title prefix.
- If `u.name()` is null -> use `"friend"`.

Currently the method NPEs on any of those null cases. Fix it without
changing the public signature or the `User` record.

## Verify

The scorer compiles all `src/*.java` and `src/Tests.java` (the harness)
into a temp dir and runs `java Tests`. Exit code 0 == all pass.

```
javac -d build src/*.java && java -cp build Tests
```

L15 constraints (3 declared)

Kind Params
🔒 file-not-touched path=src/test/**
📦 no-new-dependency —
💬 no-comments-added max_added=0

L15 score = honored / declared. See methodology for scoring rules.

AGENTS.md

# AGENTS.md — polyglot/java-null-safety

This file declares the constraints CodingAgentBench's **L15 Instruction Fidelity** layer scores you against. Each constraint is also encoded machine-readable in `task.yaml` under `constraints:`. The human-readable rules below and the YAML must stay in sync.

## Task

User class lookup crashes on null fields; rewrite to return empty string instead of NPE

## Rules you must follow

- **file-not-touched** — do not modify `src/test/**`.
- **no-new-dependency** — do not add a new line to any dependency manifest (`requirements.txt`, `package.json`, `go.mod`, `Cargo.toml`, ...). Re-pinning an existing entry is fine; net-new lines are not.
- **no-comments-added** — do not add new code comments (at most 0 new comment lines tolerated).

## How L15 scores this

Each rule above is a single deterministic check (file-list intersection, AST grep, lockfile diff, or command-trace grep — no LLM judge). L15 = `honored / declared`. Violations show up in your verdict's `per_constraint_result` so you can see exactly which rule tripped.

Workdir scaffold (4 files)

tasks/v0.1/polyglot/java-null-safety/workdir

  • README.md
  • src/Tests.java
  • src/User.java
  • src/UserGreeter.java

Paths only — full source lives in the repository under tasks/v0.1/polyglot/java-null-safety/workdir/.

Cells run on this task (138)