rill-model

Detailed instructions and examples for developing model resources in Rill

1.36x

Quality

51%

Does it follow best practices?

Impact

93%

1.36x

Average score across 3 eval scenarios

Securityby

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./skills/rill-model/SKILL.md

Quality

Content

53%Weight 40%Scale 1-5

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

The content is a strong, highly actionable Rill domain reference with excellent executable examples, but it is structured as a monolith: a 640-line JSON schema dump, a duplicated ClickHouse schema block, and the full example catalog are all inlined in SKILL.md with zero progressive disclosure and no explicit development workflow with validation steps.

Suggestions

Move the full JSON schema (and optionally the example catalog) into references/ files (e.g., references/model-schema.md, references/examples.md) and keep SKILL.md as an overview with well-signaled one-level-deep pointers.

Remove the duplicated ClickHouse output-properties block (the second allOf/then occurrence of the same ~60 lines) — it adds tokens with no new information.

Add an explicit model-development workflow with validation checkpoints: inspect available connectors, choose model category, scaffold, test against a dev partition, then verify output — so the guidance has a sequence rather than only scattered best practices.

Dimension	Reasoning	Score
Conciseness	The body is domain-dense (Rill specifics Claude does not know), but ~640 lines are a raw JSON schema dump inlined into SKILL.md, and the ClickHouse 'output' property block appears twice nearly verbatim (the allOf/then section and its repeat in the second allOf block). That duplication plus the inlined schema makes it noticeably verbose rather than 'mostly efficient with some trimming'.	2 / 5
Actionability	The skill provides ~25 complete, copy-paste-ready YAML/SQL examples covering the common cases (S3/GCS/BigQuery/Snowflake/MySQL to DuckDB or ClickHouse, partition- and state-based incremental models, dev partitions, TTL/indexes, change modes), plus a syntax scaffold and concrete property guidance. Every example is executable, not pseudocode.	5 / 5
Workflow Clarity	There is no explicit end-to-end sequence for developing a model (e.g., inspect connectors -> choose model category -> scaffold -> test with dev partitions -> validate), and no validation checkpoints for verifying a model runs or produces correct data; guidance is scattered across 'Best practices' sections instead. It stays at 3 rather than 2 because decision rules (when to materialize, which incremental strategy, dev-partition practices) are concrete and well stated.	3 / 5
Progressive Disclosure	No bundle files (references/, scripts/, assets/) exist at all; the entire 1,400-line body — including a full JSON schema reference and a 20+ example catalog that clearly belong in separate reference files — is inlined in SKILL.md. This matches 'content that clearly belongs in separate files is inlined' despite the decent section headers, which keep it above level 1.	2 / 5
	Total	12 / 20 Passed

Description

48%Weight 40%Scale 1-5

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 is domain-specific and distinctive thanks to naming Rill, but it is thin: it says nothing about when to use the skill and lists no concrete capabilities. A 'Use when...' clause with natural trigger phrases and 2-3 named actions would move it into good-examples territory.

Suggestions

Add an explicit trigger clause, e.g., 'Use when creating, editing, or debugging model resources (models.yaml / .sql files) in a Rill project, or when the user mentions Rill models, ETL, or data ingestion.'

Enumerate 2-3 concrete capabilities instead of the generic 'developing', e.g., 'Create and materialize source and derived models, configure incremental and partitioned ingestion, and set up refresh schedules.'

Include natural keyword variations users would say (SQL models, ETL, transformation, incremental models) to improve trigger-term coverage.

Dimension	Reasoning	Score
Specificity	The description names the domain ("model resources in Rill") but the only action stated is the generic "developing", padded by meta-language ("Detailed instructions and examples") that describes the document rather than a capability. It never enumerates concrete actions like materializing tables, incremental loading, or partitioning, matching the anchor 'Names the domain but actions are minimal or generic' rather than the level-3 anchor which requires 1-2 concrete actions.	2 / 5
Completeness	The 'what' is stated (developing model resources in Rill), but the 'when' is entirely absent — there is no 'Use when...' clause or equivalent trigger guidance, which the guidelines cap at 3. It is not level 2 because the 'what' is clear and domain-specific, not vague.	3 / 5
Trigger Term Quality	Relevant keywords are present ("model", "Rill", "developing") that a Rill user would naturally say, but common variations and synonyms are missing — no mention of SQL, ETL, transformation, ingestion, or terms like "create/edit a model". This fits 'Some relevant keywords but missing common variations or synonyms' rather than level 4, which expects broader natural-term coverage.	3 / 5
Distinctiveness Conflict Risk	"model resources in Rill" occupies a clear niche — the tool name makes it unlikely to trigger for unrelated skills. Minor overlap risk remains with closely related concepts (e.g., ML 'models' or sibling Rill resources like metrics views) since no explicit triggers delineate the boundary, keeping it below the level-5 anchor.	4 / 5
	Total	12 / 20 Passed

Validation

87%

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 — 14 / 16 Passed

Validation for skill structure

Criteria	Description	Result
skill_md_line_count	SKILL.md is long (1420 lines); consider splitting into references/ and linking	Warning
relative_links	Relative link issues: 7 missing, 3 suspicious	Warning

	Total	14 / 16 Passed

Repository: rilldata/agent-skills
Path: skills/rill-model/SKILL.md
Commit: 1307e7f

Reviewed: 7 days 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.