Bench
for LuumenAI
A timing harness for Luumen itself: one sleep call, no credential
Bench is the integration you connect when you want to measure Luumen, not your estate. It exposes a single read-only tool, BENCH_SLEEP, which waits for a set duration and returns. Use it to time agent round trips, check queueing behavior, or smoke-test a new install. No credential is needed, and there is nothing here that changes a live system.
The Bench toolbox
1 tools, all read-only. They answer instantly, and every call is logged.
- ReadSleepSleep
Questions
Bench needs no credential of its own, so there is nothing to authorize. Access is still granted per agent, skill, and team inside Luumen, and every call is logged.
No. Every one of the 1 tools LuumenAI has for Bench is read-only, so it can answer questions from Bench but cannot change anything in it. Reads are not gated — they answer immediately, and each one is recorded on the audit trail with the actor and the result.
You decide. Access is granted per agent, skill, and team, and per environment — production is not staging.
Every call to Bench is recorded with the actor, the input, and the result, and can be linked to the ticket or change record.
Often connected alongside
All integrations
GitHubRepos, actions, and runbook code as context and as actions.
GitLabMerge requests and pipelines proposed, triggered, and rolled back.
BitbucketRead commits, diffs, and branches in Bitbucket, then open the fix as a held write.
Azure DevOpsAzure DevOps repos and pipelines as context — and as approved actions.Put Bench to work with Luumen
Connect in minutes. Every action scoped, approved, and audited from day one.