docs: audit logs track meeting notes 2025-12-02

Signed-off-by: Alan Pinkert <apinkert@cisco.com>

Author: alanisaac

docs/contribute/wip/5_Audit_Logs_Standardization/meeting_notes/2025-12-02.md

@@ -0,0 +1,34 @@
+# Notes
+
+* Frameworks
+    * AGNTCY Observe
+        * Still needs user context
+    * OpenTelemetry
+        * How the above two are related / overlapping?
+    * OASF / OCSF
+    * Homegrown systems?
+* What vs how?
+    * Are we focused only on identity or the whole thing?
+    * User context: whether the agent is acting on behalf of a user (delegation, impersonation)
+    * Privacy?  Set aside for now, bring it in later
+* Comparison to “normal” distributed systems
+    * Deterministic vs stochastic
+    * Why was the path chosen?  What other paths could have been taken?
+* Problems we’re trying to solve
+    * What are we expecting?
+    * Capturing identity context
+    * Capturing authorization context (was it OAuth? SAML? API token?)
+
+Documents reviewed:
+* https://opentelemetry.io/docs/specs/semconv/gen-ai/gen-ai-spans/
+    * Seems like Otel’s conventions plus attributes to help record `slim`?
+* https://diforum.community/posts/trust-frameworks-for-ai-agents-traceability
+* https://github.com/agntcy/observe/tree/main/schema
+
+Action Items:
+
+* Jim: NIST controls
+* Alan: make a spreadsheet of frameworks and perform a gap assessment (check if there are already extensions)
+* Alan: create a list of problems we’re trying to solve / capabilities we expect the framework to have
+* Mohamed: Talk to Outshift folks working on Observe
+* Mohamed: Test AGNTCY observe SDK for auditing capabilities