Public assessment

jasonwilbur/oci-pricing-mcp (oci-pricing-mcp)

oci-pricing-mcp · v1.0.0 · scanned Aug 2, 2026

What changed in the harness

Selection accuracy 98→96, token cost up 12%, unconfirmed writes 100%→100%.

Category breakdown

Where the score comes from.

Earned points across the four signals Gradable measures. Safety and Legibility are scored out of 30; Economics and Discoverability are scored out of 20.

01Safety

30.0 / 30

30.0 out of 30

02Legibility

19.8 / 30

19.8 out of 30

03Economics

17.2 / 20

17.2 out of 20

04Discoverability

18.0 / 20

18.0 out of 20

Highest-impact fix

Estimated gain +1 point

Rewrite 43 underspecified tool descriptions

State each tool's behavior, explain every parameter, and declare return semantics, prioritizing tools exercised by failed selection tasks.

Description evidence

Defects and rewrites.

43 defects found across the exposed tool descriptions. Suggested rewrites make purpose, inputs, boundaries, and returns easier for an agent to understand.

Tool	Defect types	Suggested rewrite
`get_pricing`	name_restates_behavior no_return_description	Looks up OCI pricing filtered by service category and optional resource type/region. Returns matching price entries (unit price, currency, billing metric) for the specified service, or all resource types within it if 'type' is omitted.
`list_services`	no_return_description	List all OCI services with pricing categories. Filter by category: compute, storage, database, networking, kubernetes, ai-ml, observability, etc. Returns an array of service names with their associated pricing category.
`compare_regions`	no_return_description	Compare pricing for a resource across OCI regions. Note: OCI maintains consistent pricing across all commercial regions. Returns a per-region price breakdown for the given service and type, confirming price parity or highlighting any regional differences.
`list_regions`	name_restates_behavior no_return_description	Returns the full list of available OCI regions, including each region's identifier and physical location.
`calculate_monthly_cost`	name_restates_behavior params_unexplained no_return_description	Estimates total monthly OCI spend from a mixed configuration. Accepts optional 'compute' (shape, ocpus, memoryGB, hoursPerMonth), 'storage' (block/object/archive/file GB), 'database' (type, ecpus, storageGB, licenseIncluded), and 'networking' (load balancer bandwidth, outbound data GB) sections, plus an optional region; only the sections provided are priced. Returns a total estimated monthly cost with a line-item breakdown per component.
`quick_estimate`	no_return_description	Get a quick cost estimate for common deployment presets: small-web-app, medium-api-server, large-database, ml-training, kubernetes-cluster. Returns the preset's estimated monthly cost along with the underlying resource assumptions used.
`list_compute_shapes`	name_restates_behavior no_return_description	Returns OCI compute shapes (VMs, bare metal, GPU), optionally filtered by family, type, or a maximum OCPU hourly price, each with its OCPU and memory pricing.
`get_compute_shape_details`	name_restates_behavior no_return_description	Looks up a specific compute shape family and returns its detailed pricing plus available OCPU/memory configuration options.
`compare_compute_shapes`	name_restates_behavior no_return_description	Compares hourly and monthly pricing across the given list of compute shape families, returning a side-by-side price breakdown.
`list_storage_options`	no_return_description	List all OCI storage options (block, object, file, archive) with pricing. Returns each storage option's tier, unit, and price.
`calculate_storage_cost`	name_restates_behavior params_unexplained no_return_description	Estimates monthly storage cost from any combination of block volume GB (with optional performance tier: basic/balanced/high/ultra), object storage GB (with optional tier: standard/infrequent/archive), and file storage GB. Returns the total estimated monthly cost with a breakdown per storage type.
`compare_storage_tiers`	name_restates_behavior no_return_description	Compares monthly pricing across all OCI storage tiers for a given size in GB. Returns each tier's name and computed price for that size.
`list_database_options`	name_restates_behavior no_return_description	Returns OCI database offerings (Autonomous, MySQL, PostgreSQL, NoSQL, etc.), optionally filtered by database type or license type, each with its pricing.
`calculate_database_cost`	name_restates_behavior params_unexplained no_return_description	Estimates cost for a specific database configuration given its type, compute units (ECPUs/OCPUs), storage size in GB, optional license type (included or bring-your-own), and optional hours per month (defaults to full-month usage). Returns the estimated cost with a compute/storage/license breakdown.
`compare_database_options`	name_restates_behavior params_unexplained no_return_description	Compares suitable OCI database options for a given workload type (oltp, analytics, document, or general). Returns a list of matching database options with their pricing for that workload.
`list_networking_options`	no_return_description	List all OCI networking options (load balancers, FastConnect, VPN, egress) with pricing. Many networking services are FREE. Returns each networking option with its pricing or free-tier status.
`calculate_networking_cost`	name_restates_behavior params_unexplained no_return_description	Estimates monthly networking cost from any combination of flexible load balancer count and bandwidth (Mbps), network load balancer count, outbound data transfer (GB), FastConnect circuit speed (1/10/100 Gbps), VPN connection count, and NAT gateway count, applying free-tier credits where applicable. Returns the total estimated monthly cost with a per-item breakdown.
`compare_data_egress`	name_restates_behavior no_return_description	Compares OCI outbound data transfer pricing against AWS, Azure, and GCP for a given monthly data volume in GB. Returns each provider's egress cost for that volume.
`list_kubernetes_options`	name_restates_behavior no_return_description	Returns OKE (Kubernetes) cluster options and their pricing, optionally filtered by cluster type; basic clusters are free.
`calculate_kubernetes_cost`	params_unexplained no_return_description	Calculate OKE cluster cost including control plane and worker nodes. Requires cluster type (basic or enhanced), worker node count, node shape, node OCPUs, and node memory GB; an optional 'virtualNodes' object (pod OCPUs, pod memory GB, hours per month) adds virtual node pool cost. Returns the total estimated monthly cost broken down by control plane and worker/virtual node pricing.
`compare_kubernetes_providers`	params_unexplained no_return_description	Compare OKE pricing with AWS EKS, Azure AKS, and GCP GKE for a cluster with the given node count, OCPUs per node, and memory per node (GB). Returns each provider's estimated monthly cost for an equivalent cluster.
`get_free_tier`	name_restates_behavior no_return_description	Returns the list of resources and usage limits included in OCI's Always Free tier.
`get_pricing_info`	no_return_description	Get metadata about the pricing data (last updated, source, etc.). Returns the data's last-updated timestamp, source, and related metadata fields.
`list_realtime_categories`	no_return_description	List all service categories available from Oracle's real-time pricing API. Returns the list of category names usable as filters for fetch_realtime_pricing.
`list_multicloud_databases`	no_return_description	List Oracle database services available on Azure, AWS, and Google Cloud (Database@Azure, Database@AWS, Database@Google Cloud). Returns matching services with provider and database type, optionally filtered by provider or database type.
`get_multicloud_availability`	no_return_description	Get the full availability matrix showing which Oracle database products are available on which cloud providers. Returns a matrix mapping each database product to the cloud providers that offer it.
`calculate_multicloud_database_cost`	name_restates_behavior params_unexplained no_return_description	Estimates monthly cost for running an Oracle database on Azure, AWS, or Google Cloud, given the provider, database type (autonomous-serverless, autonomous-dedicated, exadata, exascale, base-db), compute units, storage GB, and optional license type (included or BYOL). Returns the estimated monthly cost with a compute/storage breakdown.
`compare_multicloud_vs_oci`	no_return_description	Compare costs of running Oracle database on OCI vs Azure/AWS/GCP. Shows price parity and billing differences. Returns per-platform cost for the given database type, compute units, and storage GB, along with parity notes.
`list_aiml_services`	no_return_description	List OCI AI/ML services including Generative AI (Cohere, Meta Llama, xAI Grok), Vision, Speech, Language, and Document Understanding. Returns matching services with their pricing, optionally filtered by type or model.
`list_observability_services`	name_restates_behavior no_return_description	Returns OCI Observability services (APM, Logging, Log Analytics, Monitoring, Stack Monitoring, etc.) with their pricing, optionally filtered by type.
`list_integration_services`	name_restates_behavior no_return_description	Returns OCI Integration services (Integration Cloud, GoldenGate, Data Integration, Streaming, Queue) with their pricing, optionally filtered by type.
`list_security_services`	name_restates_behavior no_return_description	Returns OCI Security services (Data Safe, Cloud Guard, Vault, WAF, Network Firewall, etc.) with their pricing, optionally filtered by type.
`list_analytics_services`	name_restates_behavior no_return_description	Returns OCI Analytics services (Analytics Cloud, Big Data Service, Data Flow, Data Science, etc.) with their pricing, optionally filtered by type.
`list_developer_services`	name_restates_behavior no_return_description	Returns OCI Developer services (Functions, Container Instances, API Gateway, APEX, DevOps) with their pricing, optionally filtered by type.
`list_media_services`	name_restates_behavior no_return_description	Returns OCI Media Services (Media Flow transcoding, Media Streams live/VOD) with their pricing, optionally filtered by type.
`list_vmware_services`	no_return_description	List Oracle Cloud VMware Solution (OCVS) pricing for running VMware workloads on OCI. Returns the available OCVS configurations with their pricing.
`list_edge_services`	no_return_description	List OCI Edge services including DNS, Email Delivery, and Health Checks. Returns matching services with their pricing, optionally filtered by type.
`list_governance_services`	name_restates_behavior no_return_description	Returns OCI Governance services (Access Governance, Fleet Management, License Manager) with their pricing, optionally filtered by type.
`list_exadata_services`	no_return_description	List Oracle Exadata Cloud pricing including Exascale, dedicated infrastructure, ECPUs, OCPUs, and storage. Returns matching pricing entries, optionally filtered by type.
`list_cache_services`	no_return_description	List OCI Cache with Redis pricing. Shows managed Redis cache options with low and high memory tiers. Returns each tier's configuration and price.
`list_disaster_recovery_services`	name_restates_behavior no_return_description	Returns pricing for OCI Full Stack Disaster Recovery, which automates DR across application stacks.
`list_additional_services`	no_return_description	List additional OCI services including OpenSearch, Secure Desktops, Blockchain, TimesTen, Batch, Recovery Service, ZFS Storage, Lustre, and Digital Assistant. Returns matching services with their pricing, optionally filtered by type.
`get_services_summary`	name_restates_behavior no_return_description	Returns a summary of all OCI services with pricing data, including counts by category and coverage information.

Selection evidence

Confusable tool pairs.

9 pairs where similar names or overlapping descriptions may send an agent toward the wrong tool.

Tool A	Tool B	Confidence	Why they collide
`calculate_database_cost`	`calculate_multicloud_database_cost`	high	A task like 'calculate cost for an autonomous database with 4 ECPUs and 100GB storage' doesn't specify provider, and both tools accept nearly identical computeUnits/storageGB/licenseType params, so the agent may pick the multicloud variant even for plain OCI, or vice versa when the user mentions Azure/AWS/GCP only in passing.
`get_pricing`	`get_pricing_info`	high	Names differ by only one token; 'get pricing info for compute' plausibly triggers get_pricing_info even though that tool returns metadata (last updated, source) not actual prices, while get_pricing returns real pricing data.
`list_database_options`	`compare_database_options`	medium	A request like 'what database options are best for OLTP workloads' could route to either list_database_options (type filter) or compare_database_options (workloadType enum), since both surface database choices differently.
`list_compute_shapes`	`compare_compute_shapes`	medium	'Show me pricing for E5 and A1 shapes' could be satisfied by list_compute_shapes with a family filter or misrouted to compare_compute_shapes which requires an explicit shapes array, causing selection ambiguity when the user names multiple shapes generically.
`list_compute_shapes`	`get_compute_shape_details`	medium	'Give me details on E5.Flex pricing' is ambiguous between list_compute_shapes (family filter, list of shapes) and get_compute_shape_details (single shapeFamily, detailed config options).
`calculate_monthly_cost`	`calculate_multicloud_database_cost`	low	A vague 'estimate my monthly database cost' without specifying OCI vs multicloud could plausibly go to either, though calculate_monthly_cost is broader (compute+storage+db+networking) making full confusion less likely.
`list_database_options`	`list_multicloud_databases`	medium	'List Oracle database options' is ambiguous — could mean native OCI database types (list_database_options) or Database@Azure/AWS/GCP offerings (list_multicloud_databases), especially since both share 'autonomous', 'exadata', 'base-db' type filters.
`list_services`	`list_aiml_services`	low	list_services accepts a free-text category filter including 'ai-ml', so a request like 'list AI/ML services' could go to either the generic list_services with category set or the dedicated list_aiml_services tool.
`list_edge_services`	`list_governance_services`	low	Both are niche category listers with overlapping generic description tokens ('including','list','oci','service'), but their enum types (dns/email vs access-governance/license-manager) are distinct enough that real task confusion is unlikely beyond generic browsing requests.

Compare the field

One score is useful.
The evidence makes it actionable.

Back to the leaderboard