Good AI Task

AI compatibility

AI can do the code analysis, but debugging a live memory leak still needs a human in the loop.

Possible with caveats

Workable, but read the conditions.

Average across 1 submission.

52
avg / 100

The honest read

An AI agent can meaningfully assist with heap snapshot analysis, pattern recognition, and generating patched code for common Node.js/Express/MongoDB memory leak patterns. However, the agent cannot autonomously collect live heap snapshots, run the service, or validate the fix under real production load — the most critical parts of the debugging loop require direct environment access and iterative human judgment.

Aggregated across 1 submission.

The five dimensions

Repeatability

Low

Memory leaks are highly idiosyncratic — each instance depends on the specific codebase, middleware chain, MongoDB driver version, and usage patterns. There is no repeatable structure; every diagnosis requires unique investigation.

Ambiguity Tolerance

Medium

The end goal (no crash after 48 hours, stable heap) is concrete, but confirming success requires running the patched service under real load for days — a criterion an agent cannot verify autonomously within a task window.

Data & Tool Availability

Low

The agent needs actual heap snapshots, source code, dependency versions, and ideally live profiling access — none of which are provided in the task description. Without these artifacts, the agent can only reason about common patterns, not the actual bug.

Error Cost

High

A wrong patch deployed to a service handling 10,000 requests/hour could introduce new bugs, data corruption, or worse crashes. The service is already critical and crash-prone, so a bad fix compounds the risk significantly.

Human Judgment Required

High

Identifying the true leak source from heap snapshots requires iterative hypothesis testing, domain knowledge of the specific codebase, and judgment calls about which retained objects are genuinely leaking versus expected. A human engineer's contextual knowledge of the system is hard to replace.

What an agent would need

  • Full source code of the microservice including all middleware, route handlers, and MongoDB query logic
  • Actual heap snapshot files (.heapsnapshot) captured at multiple points during the 48-hour uptime cycle
  • Dependency manifest (package.json with exact versions) to identify known leaky library versions
  • Access to a sandboxed environment or CI pipeline to run the patched service and collect profiling data
  • Load testing tooling (e.g., k6 or Artillery) configured to simulate 10,000 req/hr to validate the fix

Best-matched agent type

Code Agent

The kind of agent this work would call for if it were a fit. For this task, it isn't.

Run your own fit check

Get a calibrated read on your specific task in under a minute.

Check a task