run-evals

Run evaluations for one, multiple, or all skills using the agent orchestration framework. Make sure to use this skill whenever the user asks to run evals, test a skill's performance, run benchmarks, or compare baseline versus with-skill execution.

Quality

91%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Quality

Content

92%

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

A lean, highly actionable workflow with a clear sequenced process and explicit validation/concurrency checkpoints. The main gap is progressive disclosure: the prompt-template references point to files that are not present in the bundle.

Suggestions

Ship the referenced resources/ files (with_skill_execution_prompt.md, baseline_execution_prompt.md, agent_judge_prompt.md) so the one-level-deep references resolve and progressive disclosure is realized.

Tighten step 4's repeated confinement warnings (CRITICAL + WORKSPACE LIMITATION) into a single consolidated guard to remove mild redundancy.

Add an explicit re-grade/retry loop in the Grade step so that failed expectations feed back into a fix-and-revalidate cycle rather than only being reported.

Dimension	Reasoning	Score
Conciseness	The body is a dense, purely instructional 6-step list with no concept explanations Claude already knows; the CRITICAL/WORKSPACE LIMITATION warnings are operational guards, not padding, so most tokens earn their place.	3 / 3
Actionability	Concrete guidance throughout: specific file paths (resources/with_skill_execution_prompt.md, evals/README.md), specific commands (dart pub get, dart format, dart analyze, dart test), named profiles (bare-agent), and explicit workspace modes (Workspace: branch).	3 / 3
Workflow Clarity	A clearly numbered 1-6 sequence with an explicit Grade/validation checkpoint that requires listing both the failed expectation and what was found, plus explicit concurrency-pollution guards for risky batch operations.	3 / 3
Progressive Disclosure	The body is a clean overview with well-signaled, one-level-deep references (resources/*.md, evals/README.md), but those referenced detail files are absent from the bundle, so the disclosure structure is not actually realized and navigation would break.	2 / 3
	Total	11 / 12 Passed

Description

90%

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 with explicit what-and-when structure and natural trigger terms. Its only weakness is specificity: the capability set reads as one action with variations rather than multiple distinct concrete operations.

Dimension	Reasoning	Score
Specificity	Names the domain ('Run evaluations for one, multiple, or all skills') and the mechanism ('agent orchestration framework'), but the enumerated items (run evals, test performance, run benchmarks, compare baseline vs with-skill) are variations of one core action rather than a list of distinct concrete capabilities like the rubric's top anchor.	2 / 3
Completeness	Explicitly states what it does ('Run evaluations... using the agent orchestration framework') and when to use it ('Make sure to use this skill whenever the user asks to run evals...'), satisfying both halves with an explicit trigger clause.	3 / 3
Trigger Term Quality	Natural user phrasing is well covered: 'run evals', 'test a skill's performance', 'run benchmarks', and 'compare baseline versus with-skill execution' are terms a user would actually say.	3 / 3
Distinctiveness Conflict Risk	The skill-eval niche combined with 'agent orchestration framework' and baseline-vs-with-skill framing gives it a distinct trigger surface unlikely to fire for unrelated skills.	3 / 3
	Total	11 / 12 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
metadata_version	'metadata.version' is missing	Warning
metadata_field	'metadata' should map string keys to string values	Warning

	Total	14 / 16 Passed

Repository: flutter/agent-plugins
Path: tool/dart_skills_lint/.agents/skills/run-evals/SKILL.md
Commit: 8aaa41d

Reviewed: about 13 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.