-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathclassifier_agent.json
More file actions
52 lines (52 loc) · 2.06 KB
/
Copy pathclassifier_agent.json
File metadata and controls
52 lines (52 loc) · 2.06 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
{
"name": "FlakeWarden Triage Classifier",
"description": "Classifies an ambiguous test failure as real_defect, flaky, or environment, grounded in Test Manager evidence. Invoked by the Maestro process only for failures the deterministic scorer could not confidently bucket.",
"version": "0.1.0",
"model": {
"provider": "anthropic",
"name": "claude-opus-4-8",
"temperature": 0,
"maxTokens": 1024
},
"systemPromptFile": "classifier_prompt.txt",
"inputs": {
"type": "object",
"required": ["test_name", "stack_trace"],
"properties": {
"test_name": { "type": "string" },
"stack_trace": { "type": "string" },
"recent_dom_diff": { "type": "string" },
"commit_message": { "type": "string" },
"runner_logs": { "type": "string" }
}
},
"contextGrounding": {
"description": "Hybrid RAG over the failure's trusted artifacts. The agent retrieves, it does not free-associate.",
"sources": [
{ "name": "test-manager-results", "type": "uipath.testmanager", "scope": "failing-test-window" },
{ "name": "object-repository-diffs", "type": "uipath.objectrepository", "scope": "since-last-green" },
{ "name": "scm-commits", "type": "git", "scope": "first-red-commit" }
]
},
"outputSchema": {
"type": "object",
"required": ["label", "confidence", "rationale"],
"properties": {
"label": { "type": "string", "enum": ["real_defect", "flaky", "environment"] },
"confidence": { "type": "number", "minimum": 0, "maximum": 1 },
"rationale": { "type": "string" },
"proposed_fix": { "type": ["string", "null"] }
}
},
"guardrails": {
"default_on_uncertainty": "real_defect",
"min_confidence_for_autonomous_action": 0.55,
"never_auto_apply_fix": true,
"trust_layer": "UiPath AI Trust Layer (PII redaction + audit logging enabled)"
},
"evaluation": {
"evalSet": "../corpus/failures.jsonl",
"metrics": ["accuracy", "safety_false_positive_rate", "noise_false_alarm_rate"],
"releaseGate": { "safety_false_positive_rate": "== 0.0", "accuracy": ">= 0.88" }
}
}