Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add costory-io/costory-finops-mcp-skills --skill virtual-dimensionsgit clone --depth 1 https://github.com/costory-io/costory-finops-mcp-skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/costory-io/costory-finops-mcp-skills/virtual-dimensions)<a href="https://agentmods.dev/skills/costory-io/costory-finops-mcp-skills/virtual-dimensions"><img src="https://agentmods.dev/badge/skills/costory-io/costory-finops-mcp-skills/virtual-dimensions/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/costory-io/costory-finops-mcp-skills/virtual-dimensions"><img src="https://agentmods.dev/badge/skills/costory-io/costory-finops-mcp-skills/virtual-dimensions.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00072 | $0.04279 |
| Opus 5 | $0.00036 | $0.02139 |
| Sonnet 5 | $0.00014 | $0.00856 |
| Haiku 4.5 | $0.00007 | $0.00428 |
Grade A, and why
virtual-dimensions scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 11d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 207 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Virtual Dimensions
A virtual dimension is a custom cost axis (e.g. "Environment", "Team", "Product line") built from ordered rules. Each rule has:
conditionCel— CEL filter selecting rows (e.g.cos_environment in ["prod"])allocation— where matched spend goes. Supports all types:dimensionValue,existingColumn,splitCost,telemetry(same shapes asgetreturns for a virtual dimension)
Rules are evaluated top-to-bottom; first match wins. A leftover rule catches everything else.
Edits happen on a draft. Publishing is explicit and triggers BigQuery refresh. Production is unchanged until publish_virtual_dimension.
When to Trigger
- Building a new custom cost axis (Environment, Team, Product line, …)
- Editing rules on an existing virtual dimension
- Previewing leftover %, breakdown, or rule overlap/shadowing
- Splitting shared cost by a usage metric (
telemetryallocation) - Publishing a draft after user confirmation
Auth
Write tools (create / update / publish) are Clerk-only — not available on the service route. Read tools work on both routes.
Concepts
| Concept | Meaning |
|---|---|
| Draft vs published | Mutations touch the latest pending draft; published state stays live until publish |
| Draft persistence | preview_virtual_dimension_draft and virtual_dimension_overlap_matrix (when includeDraft: true) return draftPersisted: true when a pending draft exists, draftPersisted: false when analyzing published-only state in memory (not publishable). create_virtual_dimension_draft and update_virtual_dimension_draft persist only when validation passes — on success draftPersisted: true and draftValidation: { ok: true }; on failure nothing is saved (draftPersisted: false). publish_virtual_dimension requires a persisted pending draft — call update_virtual_dimension_draft for an existing virtualDimensionId (or create_virtual_dimension_draft for a brand-new VDIM) |
| Rule order | Earlier rules claim spend; later overlapping rules are shadowed |
| Leftovers | Spend not matched by any named rule — high leftover means add more rules |
| Leftover (MCP) | Auto catch-all returned as leftoverRule (not in rules). Do not include a catch-all in rules on create or update — duplicate leftovers cause ordering confusion. There is no leftover-label input on create/update — rename the leftover bucket in the web app |
| Rule ids | Create: saveVirtualDimensionDraft re-stamps all rule ids — response ids are authoritative. Update: preserves ids from input (including when seeding a draft from published state). Both: ids are sticky to logical rules — reorder by moving id + name + conditionCel + allocation together; attaching an id to a different condition silently mis-assigns spend |
| Tags | get returns tags (names) for both published and draft. Pass tagNames (strings) on create/update — not tag UUIDs |
values |
Output-only — derived from rule allocations; do not send on create/update |
| Allocation shapes | dimensionValue: { allocationType: "dimensionValue", dimensionValue: "prod" }; existingColumn: { allocationType: "existingColumn", existingColumn: "cos_team" }; splitCost: { allocationType: "splitCost", reAllocationParams: { type: "custom", partitions: [{ label: "A", weight: 60 }, { label: "B", weight: 40 }] } } (weights sum to 100); telemetry: { allocationType: "telemetry", externalMetric: { provider, integrationId, metricName, aggregator, groupByFields: ["<split-by>"] }, mappingType: "mapping" | "regexMapping" | "identity", mappingParams: { mapping: { "key": "value" } }, regexTransformation: "..." } — discover externalMetric via list_metrics { includeExternal: true, search }. Do not set datasource on new reallocations (leftover stored rows may still have it — echo unchanged). mapping keys are metric value names (not cos_* dimensions); unmapped values → leftover. Echo non-dimensionValue allocations from get unchanged unless the user explicitly asks to edit advanced allocation behavior |
| CEL only | Always use conditionCel strings; never raw query-builder JSON |
| Unlabelled values | Use cos_environment == null in conditionCel (not is_null or string "null") |
| Declarative updates | update_virtual_dimension_draft takes the full desired rules array, not diffs |
name vs bqName |
name is the display label (can change in drafts). bqName is the immutable BigQuery / CEL field name (e.g. virtual_environment) — set once at create_virtual_dimension_draft from the initial name and never updated, even after rename or publish. Read bqName from create/update/publish/get responses (or list_virtual_dimensions when resolving by name only) before groupBy / filterCel; never derive it from name. |
| Post-publish query | After publish, use returned bqName (not name) in query. computeStatus: REFRESHING = refresh job queued — poll get or list_virtual_dimensions until COMPLETED before querying; TO_REFRESH = draft promoted but refresh job queuing failed — retry publish_virtual_dimension or poll computeStatus until COMPLETED |
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 11d ago First seen · 207 lines · 72 tokens per session scan A b3b9b7f1ac07
virtual-dimensions is a skill published in the GitHub repository costory-io/costory-finops-mcp-skills (2 stars, last pushed 2d ago), licensed Apache-2.0. It adds 72 tokens to every session and 4,279 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
terraform-infrastructure-as-code
Comprehensive Terraform Infrastructure as Code skill covering resources, modules, state management, workspaces, providers, and advanced patterns for cloud-agnostic infrastructure deployment.
auditing-gcp-iam-permissions
Auditing Google Cloud Platform IAM permissions to identify overly permissive bindings, primitive role usage, service account key proliferation, and cross-project access risks using gcloud CLI, Policy Analyzer, and IAM Recommender.
implementing-cloud-trail-log-analysis
Implementing AWS CloudTrail log analysis for security monitoring, threat detection, and forensic investigation using Athena, CloudWatch Logs Insights, and SIEM integration to identify unauthorized access, privilege escalation, and suspicious API activity.
performing-aws-privilege-escalation-assessment
Performing authorized privilege escalation assessments in AWS environments to identify IAM misconfigurations that allow users or roles to elevate their permissions using Pacu, CloudFox, Principal Mapper, and manual IAM policy analysis techniques.
performing-cloud-log-forensics-with-athena
Uses AWS Athena to query CloudTrail, VPC Flow Logs, S3 access logs, and ALB logs for forensic investigation. Covers CREATE TABLE DDL with partition projection, forensic SQL queries for detecting unauthorized access, data exfiltration, lateral movement, and privilege escalation. Use when investigating AWS security…
securing-aws-lambda-execution-roles
Securing AWS Lambda execution roles by implementing least-privilege IAM policies, applying permission boundaries, restricting resource-based policies, using IAM Access Analyzer to validate permissions, and enforcing role scoping through SCPs.