TryCrucible — Prove your AI engineering skills

Scenario

You are given 40 small packages containing realistic security flaws such as SQL injection, hardcoded secrets, path traversal, weak authorization, and unsafe deserialization. Submit a security review, patched code, and the AI tool transcript showing how you investigated and verified the fix.

Expected output

Report identifies the vulnerability, affected file, impact, exploit scenario, and valid minimal fix

Dataset

📦

Dataset files

Scoring rubric

Clarity5%

Security report is structured and actionable

LLM usage20%

Transcript shows guided review, verification, and patch iteration rather than a single broad prompt

Robustness10%

Fix avoids introducing regressions and handles invalid or malicious inputs

Correctness25%

Vulnerability type, affected file, impact, and fix match ground truth for the test packages

Architecture20%

Patch addresses the root cause without broad rewrites or unrelated changes

Language-free evaluation

Build your solution in any language or framework — Python, TypeScript, Go, Rust, Java, C#, or anything else. The dataset artifacts may be in one language; your implementation does not need to match. TryCrucible evaluates the behaviour of your system, the quality of your AI workflow, your verification strategy, and the reproducibility of your submission — not your language choice.

Submission requirements

A public GitHub repository link
A Dockerfile in the repo root — any language or framework; the evaluator builds and runs your container
Your solution reads test_inputs.json from the working directory and writes results.json — standard I/O contract across all challenges
A decisions.md — 3–5 sentences on the key architectural and AI-workflow choices you made
The system must be fully reproducible — we clone, build, and run it against real test inputs

Evaluation contract

When you submit, the evaluator runs these steps in order:

1Clone your public GitHub repository
2Build your container from the Dockerfile in the repo root
3Mount test_inputs.json into the working directory
4Run your solution in a network-isolated sandbox (5 min limit, 512 MB RAM)
5Read results.json from the working directory
6Score correctness against hidden ground truth, then score architecture, AI workflow, robustness, and clarity

// I/O contract — same schema for every challenge

Input (provided by evaluator)

// test_inputs.json
[
  { "id": "t1", "input": { ... } },
  { "id": "t2", "input": { ... } }
]

Output (written by your solution)

Create a free account to start

Already have one? Sign in

Review and patch security vulnerabilities using AI tools