Context & Provenance
Track the origin, cryptographic provenance, taint propagation, and deterministic security boundaries of every piece of context entering an autonomous agent workflow.
Live Interception: ACTIVE
7 total
| Context ID | Source | Origin URI / Channel | Target Agent | Trust Level | Taint State | Decision | Time |
|---|---|---|---|---|---|---|---|
| ctx_001 | User Input | web_portal_form | planner_v1 | TRUSTED | CLEAN | ALLOW | 15:12:04 |
| ctx_002 | Database Query | postgresql://finance_ledger | analysis_agent_001 | HIGH | CLEAN | ALLOW | 15:14:22 |
| ctx_003 | Agent Message | orchestrator_v2 | research_agent_002 | MEDIUM | TAINTED | BLOCK | 15:16:10 |
| ctx_004 | System Policy | agentguard_policy_engine | orchestrator_v2 | TRUSTED | CLEAN | ALLOW | 15:16:45 |
| ctx_005 | External MCP | mcp://external-research.io/search | orchestrator_v2 | UNTRUSTED | TAINTED | BLOCK | 15:17:02 |
| ctx_006 | HTTP Webhook | https://api.thirdparty-feed.com/v1 | crawler_bot | MEDIUM | CLEAN | MONITOR | 15:18:30 |
| ctx_007 | RAG Vector Store | qdrant://compliance_kb | analysis_agent_001 | TRUSTED | CLEAN | ALLOW | 15:19:15 |