Issues

• No check compares the monitoring subnet's security group with its configured rules

  enhancement
  sindredg/ai-security-platform#178 · Filter by author sindredg opened on Oct 24, 2026

• The orchestrator's footprint during a job has not been measured

  enhancement
  sindredg/ai-security-platform#176 · Filter by author sindredg opened on Oct 24, 2026

• The daily cap has 2 % to spare on a day with three jobs

  enhancement
  sindredg/ai-security-platform#175 · Filter by author sindredg opened on Oct 24, 2026

• Stopping a workflow with argo terminate has not been tried

  enhancement
  sindredg/ai-security-platform#174 · Filter by author sindredg opened on Oct 24, 2026

• Eight checks pass when their command fails, because an empty answer counts as zero

  bug
  sindredg/ai-security-platform#173 · Filter by author sindredg opened on Oct 24, 2026

• No pod has a limit on its number of processes

  enhancement
  sindredg/ai-security-platform#166 · Filter by author sindredg opened on Oct 23, 2026

• The journal pull's own logins make up a third of the workspace's daily volume

  enhancement
  sindredg/ai-security-platform#164 · Filter by author sindredg opened on Oct 23, 2026

• The workflow diagram does not show the stop for the operator

  documentation
  sindredg/ai-security-platform#153 · Filter by author sindredg opened on Oct 22, 2026

• The evidence tool hands the model the stored file as it is

  enhancement
  sindredg/ai-security-platform#152 · Filter by author sindredg opened on Oct 22, 2026

• The model calls a token whose grants equal its contract confirmed

  enhancement
  sindredg/ai-security-platform#151 · Filter by author sindredg opened on Oct 22, 2026

• fwupd-refresh fails on the router and the edge: the firmware metadata host times out

  bug
  sindredg/ai-security-platform#145 · Filter by author sindredg opened on Oct 21, 2026

• Goose offers its own load_skill tool and two bundled skills to every job

  enhancement
  sindredg/ai-security-platform#139 · Filter by author sindredg opened on Oct 20, 2026

• Microsoft.PolicyInsights is registered outside Terraform, so the policy compliance view had no data

  enhancement
  sindredg/ai-security-platform#133 · Filter by author sindredg opened on Oct 20, 2026
  1

• No linter reads the systemd unit files

  enhancement
  sindredg/ai-security-platform#131 · Filter by author sindredg opened on Oct 19, 2026
  1

• kube-bench 5.1.6 fails while its own audit finds every pod compliant or exempt

  enhancement
  sindredg/ai-security-platform#127 · Filter by author sindredg opened on Oct 19, 2026
  1

• The network foundation runbook's target steps have not run in their current form

  enhancement
  sindredg/ai-security-platform#108 · Filter by author sindredg opened on Oct 17, 2026

• The signature check has not been tried on a pull request from a fork

  enhancement
  sindredg/ai-security-platform#95 · Filter by author sindredg opened on Oct 17, 2026

• M.5 has never been seen to fail

  enhancement
  sindredg/ai-security-platform#64 · Filter by author sindredg opened on Oct 15, 2026
  1

• T.3 to T.5 have only been seen to fail on absence, not on a changed field

  enhancement
  sindredg/ai-security-platform#63 · Filter by author sindredg opened on Oct 15, 2026
  1

• Dependabot on tool-service/ is unproven with hashed requirement files

  enhancement
  sindredg/ai-security-platform#46 · Filter by author sindredg opened on Oct 13, 2026
  1

• Evidence is stored per target, so two jobs on one target read the same evidence

  enhancement
  sindredg/ai-security-platform#45 · Filter by author sindredg opened on Oct 13, 2026

• Measure thinking again on the first real investigation

  enhancement
  sindredg/ai-security-platform#36 · Filter by author sindredg opened on Oct 12, 2026

• The deployment certificate rotation procedure has never been run

  enhancement
  sindredg/ai-security-platform#27 · Filter by author sindredg opened on Oct 12, 2026

• The cleanup after a failed model build is tested with stubs only

  bug
  sindredg/ai-security-platform#26 · Filter by author sindredg opened on Oct 12, 2026

• Checks C.2 and C.3 prove nothing about the deny rule until a target exists

  enhancement
  sindredg/ai-security-platform#23 · Filter by author sindredg opened on Oct 12, 2026