SimranShaikh commited on
Commit
03cd3fb
·
verified ·
1 Parent(s): 85750fc
Files changed (1) hide show
  1. openenv.yaml +6 -77
openenv.yaml CHANGED
@@ -1,77 +1,6 @@
1
- name: code-review-env
2
- version: "1.0.0"
3
- description: >
4
- An OpenEnv reinforcement-learning environment where AI agents learn to
5
- review source code for syntax errors, logic bugs, and security vulnerabilities.
6
- Agents receive partial reward signals at each step, enabling curriculum learning
7
- from easy syntax checks up to multi-vulnerability security audits.
8
-
9
- author: Simran Shaikh
10
- tags:
11
- - openenv
12
- - code-review
13
- - security
14
- - software-engineering
15
- - real-world
16
-
17
- language: python
18
-
19
- tasks:
20
- - id: easy_syntax
21
- name: "Python Syntax Error Detection"
22
- difficulty: easy
23
- max_steps: 5
24
- description: >
25
- Identify and fix a syntax error hidden in a short Python function.
26
- Graded on: correct issue_type, description accuracy, and fix validity.
27
-
28
- - id: medium_logic
29
- name: "Logic Bug: Off-by-One in Palindrome Check"
30
- difficulty: medium
31
- max_steps: 8
32
- description: >
33
- Find and fix a subtle off-by-one indexing bug in a palindrome checker.
34
- Fix is verified by executing 5 deterministic test cases.
35
-
36
- - id: hard_security
37
- name: "Security Vulnerability: SQL Injection, Path Traversal & Weak Hashing"
38
- difficulty: hard
39
- max_steps: 10
40
- description: >
41
- Perform a full security audit of an authentication module containing
42
- three distinct vulnerabilities. All must be identified and remediated.
43
-
44
- api:
45
- reset: POST /reset
46
- step: POST /step
47
- state: GET /state
48
- tasks: GET /tasks
49
-
50
- reward:
51
- min: 0.0
52
- max: 1.0
53
- type: continuous
54
- partial_credit: true
55
-
56
- observation_space:
57
- task_id: string
58
- task_name: string
59
- difficulty: string
60
- language: string
61
- code_snippet: string
62
- context: string
63
- step_number: integer
64
- max_steps: integer
65
- previous_feedback: string | null
66
-
67
- action_space:
68
- identified_issues:
69
- type: array
70
- items:
71
- line_number: integer | null
72
- issue_type: "syntax_error | logic_bug | security_vulnerability | performance | style"
73
- description: string
74
- severity: "low | medium | high | critical"
75
- suggested_fix: string | null
76
- explanation: string
77
- done: boolean
 
1
+ spec_version: 1
2
+ name: code_review_env
3
+ type: space
4
+ runtime: fastapi
5
+ app: server.app:app
6
+ port: 7860