ai-models

Research, compare, and update shared AI model JSON for TypeScript, web, and Rust consumers. Covers text model tiers, image and video generation models, image tool models, release provenance, pricing data sourcing, and provider-cost metering against prepaid org credit. Use when bumping model versions, adding new models, updating pricing, or auditing model specs against provider documentation.

Quality

87%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Quality

Content

82%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.

A dense, highly actionable workflow document: every common operation (add model, update pricing, regenerate projections) has executable commands and concrete data shapes, and validation is enforced via a check gate and a comprehensive final checklist. Weaknesses are modest — some rhetorical padding, no explicit error-recovery loop, and a monolithic structure that inlines detail which could live in reference files.

Suggestions

Add an explicit error-recovery loop to the generation workflow (e.g., "If generate.mjs --bundle --check fails: regenerate, rebuild both packages, then re-run the check") to turn the implied feedback cycle into an instructed one.

Move the per-domain authoring detail (Video Models card shape and pricing tables, release-provenance source-priority rules) into references/ files one level deep, keeping SKILL.md as a leaner overview with clearly signaled pointers.

Trim rhetorical policy statements (e.g., "Its shape must never be a record of how recently someone got round to updating it", "keeping it is not caution") to single-line imperatives so every token carries instruction.

Dimension	Reasoning	Score
Conciseness	Nearly all content is project-specific knowledge (canonical video cards, generator pipeline, provenance contract, deprecation policy) that Claude cannot know otherwise, but a few rhetorical policy flourishes ("a stale entry is a wrong answer, not a conservative one"; "keeping it is not caution") and repeated provider-binding rules across sections could be trimmed.	4 / 5
Actionability	Copy-paste-ready commands throughout: the five-step generate.mjs/pnpm sequence, model_info.py invocations with flags, "pnpm --filter @grida/ai-models test" and "cargo test -p grida-ai --locked", plus concrete JSON card shapes, the ImageModelPricing discriminated union, and per-model field lists covering the common update cases.	5 / 5
Workflow Clarity	The edit-JSON → generate → build → bundle → check sequence is clearly laid out with purposes explained, and the "After Any Update" checklist supplies eleven explicit validation items including test commands and the --bundle --check gate. Falls short of a 5 because error recovery is implied ("Builds/typechecks reject stale source projections") rather than given as an explicit validate → fix → re-run loop.	4 / 5
Progressive Disclosure	Well-organized sections, a Key Files table mapping every referenced repo path, and the single bundle script (scripts/model_info.py) referenced correctly and present in the bundle. Minor gap: the 354-line body inlines per-domain authoring detail (video/image card shapes, provenance rules, provider tables) that could be split into references/ files to slim the main file.	4 / 5
	Total	17 / 20 Passed

Description

92%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.

A strong description: concrete actions, comprehensive scope enumeration, an explicit "Use when" trigger clause, and a distinct niche. Third person voice is used throughout with no fluff. Only minor trigger-synonym coverage is missing.

Dimension	Reasoning	Score
Specificity	"Research, compare, and update shared AI model JSON" names multiple concrete actions, and the enumerated scope ("text model tiers, image and video generation models, image tool models, release provenance, pricing data sourcing, and provider-cost metering against prepaid org credit") comprehensively covers the catalogue domains.	5 / 5
Completeness	Explicitly answers both "what" (research, compare, and update shared AI model JSON with enumerated coverage areas) and "when" (an explicit "Use when..." clause listing four concrete triggers).	5 / 5
Trigger Term Quality	"Use when bumping model versions, adding new models, updating pricing, or auditing model specs against provider documentation" provides good natural trigger coverage, but misses common variations users might say such as deprecating/removing models or updating the model catalog. Fits between the 4 and 5 anchors: solid coverage with a few natural terms missing.	4 / 5
Distinctiveness Conflict Risk	The niche is clearly distinct — shared AI-model catalogue JSON consumed by TypeScript, web, and Rust, with release provenance and prepaid org-credit metering — so trigger overlap with other skills is minimal.	5 / 5
	Total	19 / 20 Passed

Validation

93%

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 — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
relative_links	Relative link issues: 1 suspicious	Warning

	Total	15 / 16 Passed

Repository: gridaco/grida
Path: .agents/skills/ai-models/SKILL.md
Commit: d086528

Reviewed: 4 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.