rill-development

Overview of how to develop a Rill project

1.08x

Quality

—

Does it follow best practices?

Impact

97%

1.08x

Average score across 2 eval scenarios

Securityby

Passed

No known issues

Quality

Content

77%

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The content is highly actionable with concrete properties, commands, tool guidance, and a clearly sequenced workflow with validation checkpoints. Its main weakness is a monolithic structure that inlines detailed per-resource descriptions instead of delegating them to one-level-deep reference files, slightly hurting conciseness and progressive disclosure.

Suggestions

Move the per-resource-type detail sections (Models, Metrics views, Explores, etc.) into separate reference files and keep SKILL.md as a lean overview with one-level-deep links, improving both conciseness and progressive_disclosure.

Trim background exposition like the Rill principles list and platform introductions that a data-engineering agent can infer, keeping only Rill-specific conventions.

Make the references to "separate resource-specific instruction files" into explicit, named links (e.g. [MODELS.md](references/models.md)) so navigation is concrete rather than implied.

Dimension	Reasoning	Score
Conciseness	The body is mostly efficient and well-structured, but sections like the Rill principles list and the verbose introductions to each resource type include background context (e.g. "Rill is a business intelligence platform built around the following principles") that a competent agent largely knows or could infer.	2 / 3
Actionability	It gives concrete, executable guidance: real file names (rill.yaml, .env), specific properties (olap_connector:, managed: true, mode: read), CLI commands (rill validate), tool names with usage caveats, and concrete pitfalls like "Always create new models as .yaml files, not .sql files".	3 / 3
Workflow Clarity	The "Recommended workflow" is an explicitly sequenced 8-step pipeline with a validation/iteration checkpoint ("Check for errors and keep iterating until they are fixed"), plus a fallback flow when MCP tools are unavailable and a loading-documentation prerequisite step.	3 / 3
Progressive Disclosure	The body is a single ~280-line monolith covering all resource types inline rather than a lean overview pointing to per-resource reference files, although it repeatedly signals deeper docs ("find detailed descriptions ... in the separate resource-specific instruction files", docs.rilldata.com, rill-metrics-view skill).	2 / 3
	Total	10 / 12 Passed

Description

35%

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

The description identifies the domain (Rill BI platform development) but is vague about concrete actions and omits explicit "Use when" trigger guidance. It is distinguishable mostly by the Rill proper noun rather than by rich trigger terms.

Suggestions

Add concrete capabilities, e.g. "Configure connectors, build models, define metrics views, and create dashboards in Rill."

Append an explicit trigger clause, e.g. "Use when developing or debugging a Rill BI project, or when the user mentions Rill, metrics views, or Rill dashboards."

Include natural trigger terms users would say ("Rill project", "Rill dashboard", "metrics layer", "explore dashboard") to improve trigger_term_quality.

Dimension	Reasoning	Score
Specificity	The phrase "develop a Rill project" names the domain and a high-level action but does not enumerate concrete actions like configuring connectors, modeling data, or building dashboards.	2 / 3
Completeness	It states what the skill covers (developing a Rill project) but provides no "Use when..." clause or explicit trigger guidance, so the "when" is only implied and completeness is capped at 2 per the guidelines.	2 / 3
Trigger Term Quality	"Overview of how to develop a Rill project" offers no natural trigger keywords a user would say; it relies on the proper noun "Rill" and lacks common variations like "Rill dashboard", "metrics view", or "Rill project".	1 / 3
Distinctiveness Conflict Risk	The proper noun "Rill" gives it a distinct niche unlikely to conflict with unrelated skills, but the generic "develop a project" framing could overlap with general BI or data-engineering skills.	2 / 3
	Total	7 / 12 Passed

Validation

100%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: rilldata/agent-skills
Commit: 892996d

Reviewed: about 8 hours ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.