From ac644526e3050a025af7183f08d5f431b09df177 Mon Sep 17 00:00:00 2001
From: Gautam Bharti <3028386+gautamgb@users.noreply.github.com>
Date: Mon, 14 Sep 2026 19:25:37 -0700
Subject: [PATCH] data_validation: point the agentdataflow central schema at
 the dataset-local one

#62 settled the direction: dataset-local schemas are authoritative, the
central data_validation/schemas/*.json are historical stubs. This is the
follow-up promised there for agentdataflow_trace.schema.json specifically.

The file becomes a $ref into
datasets/agentdataflow_toolexchange_traces/schema.json, with a $comment
naming the concrete divergences (trace_id shape, typed spans instead of an
untyped array, the disposition enum, required provenance evidence) and
pointing at #62 for why each one is deliberate. Nothing in the codebase
currently loads or validates against this file - grepped for both
"agentdataflow_trace.schema.json" and "exploit.schema.json" across .py,
.md, and .yml and found no reference - so this is a documentation fix, not
a behavior change, and there is nothing to test beyond the file parsing as
JSON, which it does.

The same drift exists between datasets/exploit_dataset/schema.json and
data_validation/schemas/exploit.schema.json; filed as a separate issue
rather than fixed here, since it's a different dataset's call to make and
this PR should stay reviewable as one change.
---
 .../schemas/agentdataflow_trace.schema.json     | 17 +++--------------
 1 file changed, 3 insertions(+), 14 deletions(-)

diff --git a/data_validation/schemas/agentdataflow_trace.schema.json b/data_validation/schemas/agentdataflow_trace.schema.json
index e2786fd..361a6dc 100644
--- a/data_validation/schemas/agentdataflow_trace.schema.json
+++ b/data_validation/schemas/agentdataflow_trace.schema.json
@@ -1,16 +1,5 @@
 {
-  "$schema": "http://json-schema.org/draft-07/schema#",
-  "title": "Agent Data Flow and Tool Exchange Trace",
-  "type": "object",
-  "required": ["trace_id", "category", "dsgai_mapping", "trace_data", "type"],
-  "properties": {
-    "trace_id": { "type": "string", "pattern": "^TRACE-\\d{4}$" },
-    "category": { "type": "string", "enum": ["tool_call", "multi_agent_delegation", "plugin_data_exchange", "credential_flow", "context_accumulation", "memory_read_write"] },
-    "dsgai_mapping": { "type": "array", "items": { "type": "string", "pattern": "^DSGAI(0[1-9]|1[0-9]|2[01])$" }, "minItems": 1 },
-    "agent_framework": { "type": "string" },
-    "trace_data": { "type": "array", "items": { "type": "object" }, "minItems": 1 },
-    "sensitivity_annotations": { "type": "array", "items": { "type": "string" } },
-    "security_observations": { "type": "string" },
-    "type": { "type": "string", "enum": ["benign", "adversarial"] }
-  }
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$comment": "Historical stub, superseded. datasets/agentdataflow_toolexchange_traces/schema.json is authoritative for this dataset: it is the schema its entries are actually validated against (datasets/agentdataflow_toolexchange_traces/validate.py), and it diverged from this file deliberately - trace_id shape, typed spans instead of an untyped trace_data array, a disposition enum that admits an unintentional failure alongside benign/adversarial, and a provenance tier with required evidence. See PR #62 for the reasoning behind each divergence. This file is kept only so a $ref into it does not break; do not hand-edit it without also updating the dataset-local schema, and prefer editing the dataset-local one.",
+  "$ref": "../../datasets/agentdataflow_toolexchange_traces/schema.json"
 }