Evaluate autonomous agent run infrastructure health against a ten-dimension rubric.
npx skills add https://github.com/EdgeApp/edge-dev-agents --skill orch-eval
Please help me install this Agent Skill. Skill: orch-eval Source: https://github.com/EdgeApp/edge-dev-agents/tree/main/.cursor/skills/orch-eval Command: npx skills add https://github.com/EdgeApp/edge-dev-agents --skill orch-eval
This Skill addresses the difficulty of auditing autonomous agent runs by providing a standardized, evidence-based framework to evaluate infrastructure health and operational performance.
Run the orch-eval skill by providing the specific Asana GID for the agent run you wish to audit.
High-intent search queries and answers about installing and using this skill.
You can audit autonomous agent infrastructure health by evaluating logs, manifests, and system signals against a predefined ten-dimension rubric to grade orchestration stability and generate a comprehensive audit report.
Orchestration stability grading evaluates autonomous agent runs across ten specific dimensions, including resource release, memory pressure, and slot citizenship, to distinguish successful operations from infrastructure bugs and captured errors.
You verify environment cleanup and detect fork-storms by running an infrastructure audit using local forensics, memory monitors, and watchdog logs to check if the agent adhered to the workspace status contract.
Yes, evaluating agent run health requires access to local forensics, memory monitors, and watchdog logs to analyze operational footprints and generate a comprehensive evidence-based audit report.
Yes, you can use this audit tool after an automated Asana-to-PR workflow by providing the specific Asana GID to verify the agent run cleaned up its environment and adhered to workspace contracts.
The rubric-based evaluation requires complete access to local forensics and watchdog logs, meaning runs without proper manifest logging or system signal capture cannot be comprehensively graded across all ten dimensions.
winsorllcRun scheduled diagnostics and monitor CPU, memory, and disk usage.
webrix-aiCorrelate CVE scan results with infrastructure to prioritize patching and create Jira tickets.
proofoftomCoordinate autonomous agents via A2A protocol with Agent Card discovery and task lifecycles.
paulinghamVerify post-deploy health and trigger rollback on failed checks.
