Add staged discovery routing with root and scoped traversal phases to infrastructure providers.
npx skills add https://github.com/mondoohq/mql --skill staged-discovery
Please help me install this Agent Skill. Skill: staged-discovery Source: https://github.com/mondoohq/mql/tree/main/.claude/skills/staged-discovery Command: npx skills add https://github.com/mondoohq/mql --skill staged-discovery
Staged discovery solves the problem of providers loading too much data and exhausting memory or slowing down discovery by splitting traversal into scoped phases.
Ask the AI to add staged discovery support to your provider by introducing an OptionStagedDiscovery router in Discover(), implementing a root stage that emits child scope assets, and implementing one or more scoped stages that return leaf resources with correct connection config behavior and discovery-target platform ID filtering.
High-intent search queries and answers about installing and using this skill.
Memory optimization during provider discovery is achieved by splitting traversal into a root stage and subsequent scoped stages, ensuring the MQL resource cache is released when a scope closes. This prevents loading too much data and exhausting memory.
Phased asset traversal breaks discovery into a root stage and subsequent scoped stages so traversal happens one scope level at a time. This approach prevents providers from loading too much data and slowing down discovery.
Implement staged discovery routing by introducing an OptionStagedDiscovery router in Discover(), creating a root stage that emits child scope assets, and implementing scoped stages that return leaf resources with correct connection config behavior and platform ID filtering.
Staged discovery is designed for providers with hierarchical resource trees such as AWS, GCP, Azure, and Kubernetes. It optimizes traversal performance and memory usage by processing scoped phases one level at a time.
Single-pass Discover() is preserved to maintain legacy compatibility while the new OptionStagedDiscovery router handles phased traversal. This ensures existing discovery flows remain unaffected when staged traversal is introduced.
Filter scannable assets by stripping platform IDs from non-targeted intermediate scopes while still emitting them so traversal can continue. This ensures only targeted leaf resources within a scope use parent connection linkage for API cache reuse.
n4vyshDiscover cloud resources with Terraform Search queries and generate bulk import configurations.
ntk148vAnalyze ClickHouse cache systems and produce a prioritized tuning plan.
discord-phpCoordinate DiscordPHP infrastructure utilities for caching, endpoint binding, and exceptions.
samberDiagnose Golang performance bottlenecks with pprof and benchstat benchmarks.
