LSOS

LeeromOS Console

Triage from the board — stuck reason & recommendation, inline
DashboardContactsCalendarMessagingMailTasksGenomeToolsHealthAdmin Console
BoardQueueOpsSweepsSkill AtlasPipelineTour
LSOS

LeeromOS Console

Pipeline — commands, skills, hooks, crons & the quality gate
DashboardContactsCalendarMessagingMailTasksGenomeToolsHealthAdmin Console
BoardQueueOpsSweepsSkill AtlasPipelineTour
LSOS

LeeromOS Console

Pipeline — commands, skills, hooks, crons & the quality gate
DashboardContactsCalendarMessagingMailTasksGenomeToolsHealthAdmin Console
BoardQueue2OpsSweepsSkill AtlasPipelineTour
Nodes
Edges
Hooks
Fail-closed
Gate layers
Commands

Computed from harness/ · captured 2026-08-08 14:51 UTC · structural edges never cached

Workflows
3
CI steps
30
Gate steps
8
CI·onpushpull_request·.github/workflows/ci.yml
1. Deterministic tests — Logic errors, regressions1. Deterministictests2. Static validation — Type errors, lint violations2. Static validation3. Schema validation — Malformed outputs, drift3. Schema validation4. Regression replay — Previously-fixed bugs returning4. Regression replay5. Code-graded evals — Automated rubric scoring5. Code-graded evals6. Model-graded evals — Quality beyond deterministic grading6. Model-gradedevals7. Citation/evidence verification — Ungrounded claims7. Citation/evidenceverification8. Screenshot-based review — Visual/dark-mode/mobile regressions8. Screenshot-basedreview9. Adversarial review — Hidden assumptions, security, arch9. Adversarialreview10. Human-graded evals — Taste, ambiguous UX judgment10. Human-gradedevals11. Prompt-injection / memory poisoning — Sentinel security checks11. Prompt-injection/ memory poisoning12. Performance benchmarking — Latency/token-budget regressions12. PerformancebenchmarkingStep 1: actions/checkout@v4actions/checkout@v4Step 2: Guard - no unbounded scanner subprocessGuard - no unboundedscanner subprocess

Select a CI step or a gate layer to see what it runs and which quality gate it satisfies.