Tessl • The management layer for AI agent skills

SkillPublic

eval-skills

Audit a project or library of skills for metadata, trigger boundaries, workflow completion criteria, local resource closure, instruction economy, tool scope, effort, and routing evidence. Use before shipping skills, after importing a skill collection, or when a skill triggers too often or not at all.

Quality

86%

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

77%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 body is highly actionable with an exemplary done-when checkpoint structure, but it is a monolithic reference document: version-stamped details and branch-only host references sit inline, hurting both token economy and progressive disclosure. Splitting the frontmatter/substitution/budget reference material into separate files would address the two lowest dimensions at once.

Suggestions

Move the Claude Code frontmatter field table, string-substitution table, and listing-budget sections into references/ files (e.g. references/frontmatter.md) loaded only when auditing Claude Code native skills, keeping the body to the discovery, scoring, and reporting workflow.

Consolidate scattered version-stamped details (v2.1.196, v2.1.218, v2.1.233, v2.1.252, v2.1.284) into a single version-requirements note or a dedicated compatibility section so they age in one place.

The effort-level inference guide and routing-evidence methodology are only needed for specific criteria; consider compressing them to one-line pointers with the detail behind a reference file.

Dimension	Reasoning	Score
Conciseness	The body is dense and expert-level with no padding, but it scatters time-sensitive version stamps inline ("v2.1.218", "v2.1.233", "v2.1.252", "v2.1.284") with no deprecated/old-patterns section, and inlines roughly 60 lines of branch-only frontmatter reference material that every invocation pays for. It is mostly efficient but could be meaningfully tightened.	3 / 5
Actionability	Guidance is fully executable: copy-paste discovery commands ("find .claude/skills .agents/skills -name SKILL.md -not -path '/.trash/'"), concrete validation ("claude plugin validate <skills-dir>", "skillOverrides set to 'off'"), an exact 25-point scoring table, and a complete report template covering the common output cases.	5 / 5
Workflow Clarity	The four execution steps each end in an observable "Done when:" checkpoint, parse failures are confirmed with a validation command, and the report step enforces count reconciliation — explicit validation checkpoints with a defined termination condition.	5 / 5
Progressive Disclosure	The single ~290-line file is well-sectioned but contains no references at all, and large clearly separable blocks (the Claude Code frontmatter field table, substitution placeholders, listing budgets) are inlined into the always-loaded body instead of being split into branch-loaded reference files.	3 / 5
	Total	16 / 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: it states a precise capability list, gives explicit and concrete use-when guidance, and occupies a distinct niche. The only weakness is slightly thin synonym coverage of natural trigger phrasings, which the separate when_to_use field partially compensates for.

Dimension	Reasoning	Score
Specificity	The description enumerates eight concrete audit criteria ("metadata, trigger boundaries, workflow completion criteria, local resource closure, instruction economy, tool scope, effort, and routing evidence"), giving comprehensive coverage of the skill's capability with no obvious gap.	5 / 5
Completeness	It explicitly answers both what ("Audit a project or library of skills for…") and when ("Use before shipping skills, after importing a skill collection, or when a skill triggers too often or not at all") with concrete trigger guidance.	5 / 5
Trigger Term Quality	Natural phrases like "shipping skills", "importing a skill collection", and "a skill triggers too often or not at all" are present, but common variations users would say ("review", "check", "evaluate", "score", "quality") are absent, leaving a few natural terms missing rather than comprehensive synonym coverage.	4 / 5
Distinctiveness Conflict Risk	The skill-auditing niche is clear and its triggers (pre-ship review, post-import audit, mis-triggering skills) are distinct enough that no closely related skill would claim the same prompts, keeping conflict risk minimal.	5 / 5
	Total	19 / 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
allowed_tools_field	'allowed-tools' contains unusual tool name(s)	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	14 / 16 Passed

Repository: FlorianBruniaux/claude-code-ultimate-guide
Path: examples/skills/eval-skills/SKILL.md
Revision: e0d8cbb

Reviewed: 10 days ago

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.

eval-skills

Table of Contents

Is this your skill?