Repository navigation
[uk ai resilience] UK AI Open Code Risk & Resilience Governance – 2026-11-11 #76892
Closed
Replies: 1 comment
|
This discussion has been marked as outdated by UK AI Operational Resilience. A newer discussion is available at Discussion #77186. |
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Executive summary
Lookback 7 days: 370 commits (201 with security signals), 26 open security issues, 23 open code-scanning alerts, 0 secret-scanning alerts. Sub-agents were not run and no deep control verification was done, so confidence is partial. Findings come only from the alert metadata.
Asset graph summary (recent-change scoped)
actions/setup/sh/(MCP gateway, CLI/DIFC proxy, git patch, usage artifact collection, cache memory commit). These run in the runtime setup path.actions/setup/js/(mcp_handler_shell.cjs,mcp_http_server_runner.cjs).pkg/cli/add_package_ownership.go,mcp_logs_guardrail.go).upgrade-test.yml,copilot-maintenance.yml,spec-librarian,smoke-copilot,jsweep).install-gh-aw.sh.Tier classification
actions/setup/sh/*actions/setup/jsMCP handlersupgrade-test.yml,copilot-maintenance.yml)npm/uv pipinstalls (supply chain)pkg/cliGoNo Tier C/D candidates were identified.
Control verification gaps
Risk scoring
Qualitative only. Exposure amplification is high for setup scripts because they run in every compiled workflow. Patchability is high, since the fixes are small and local. Ownership confidence is unknown.
Remediation queue
actions/setup/shnpm ci,uv pip --require-hashes) in workflowsHuman review is triggered for any change to the setup scripts and the MCP gateway.
Exception register
None.
Operational metrics baseline
Issues
No issues were created because issues are disabled on github/gh-aw.
All reactions