Acting as operations@arctransformationgroup.com (configured, not authenticated)

API http://127.0.0.1:8787 ok v2.1.0

Window: the most recent 8 evidence records, 2026-10-26 09:18:23Z to 2026-10-26 09:18:30Z. Up to 200 are read, so a busier deployment’s window is shorter in time. Generated 2026-10-26 09:18:30Z. Pending approvals and the inventory counts below are current, not windowed.

Decision outcomes

Outcome mix within the window
allowed executed pending approval approved rejected expired denied failed
0 (0%) 3 (38%) 2 (25%) 0 (0%) 0 (0%) 0 (0%) 3 (38%) 0 (0%)

Policy activity

Decisions per policy and rule within the window
Policy / rule ALLOW APPROVAL DENY
FINANCE-004/needs-approval 0 3 0
FINANCE-004/over-ceiling 0 0 1
GITHUB-002/no-default-branch 0 0 1
GITHUB-002/issues-and-prs 1 0 0
engine default 0 0 1
FINANCE-004/under-500 1 0 0

Tool activity and execution

Execution per tool within the window
Tool Runs Failures Median ms Max ms
stripe.refund 2 0 1 1
stripe.get_charge 0 0 — —
github.create_issue 1 0 0 0

A failed execution is a failed execution, not a denial: the decision allowed it and the connector could not carry it out. Both are in Evidence.

Approvals

Pending is current. The resolved counts are within the evidence window, because the API lists pending requests only — there is no approval-history endpoint to count against.

1 pending now1 approved in window0 rejected in window0 expired in window2 held for approval in window

Inventory (current)

2 agents0 suspended3 tools1 critical-tier tools2 policies47 audit records

Connector health
Connector Health Detail (as reported)
simulation healthy 3 simulated tool(s)

Audit head sequence 47. This page computes nothing the records do not already say.