agent-instructions

Design, audit, and validate repository guidance across AGENTS.md, CLAUDE.md, and `.agents/skills`. Use when creating skills, tuning descriptions, or reviewing instruction sprawl and routing.

Quality

73%

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

Fix and improve this skill with Tessl

tessl review fix ./.agents/skills/agent-instructions/SKILL.md

Quality

Content

71%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 well-structured, opinionated instruction skill with strong progressive disclosure and actionable, validated workflows. Its main weakness is conciseness: the body is long and over-explains principles Claude already knows, working against its own stated line budget.

Suggestions

Trim the multi-paragraph essays in 'Guide Without Over-Steering' and 'Write The Description First' to the rule-plus-reason pattern the skill itself prescribes, pushing toward its own <100 line target.

Convert the core design workflow ('Design The Skill From The Interaction') into an explicitly sequenced loop with a validation/retry checkpoint so workflow clarity reaches 5.

Move the long CLI-validation prose in 'Validate With Vercel CLI' into a reference or a single example block to reduce in-body token cost.

Dimension	Reasoning	Score
Conciseness	At ~370 lines the body exceeds its own 'under 100 lines when practical' guidance and repeatedly explains principles Claude already knows in long prose essays, matching 'mostly efficient but includes some unnecessary explanation'; not 2 because the prose is tighter and less padded than the verbose anchor.	3 / 5
Actionability	Concrete Vercel CLI commands, bun script invocations, symlink steps, a YAML frontmatter example, and a six-step eval loop provide mostly executable guidance; not 5 because much of the guidance is principle-driven prose rather than copy-paste ready steps.	4 / 5
Workflow Clarity	Sequences like Create A Skill, Validate With Vercel CLI, and the Evaluate loop carry clear ordering with validation checkpoints ('audit-skill-links.ts reports no dead link', 'Validation passed'); not 5 because the core authoring workflows are prose-mediated without explicit fix-retry feedback loops.	4 / 5
Progressive Disclosure	SKILL.md is an overview with three well-signaled, one-level-deep references each carrying a concrete load condition, and all referenced files (references/.md, scripts/.ts) exist, matching the 'clear overview with well-signaled one-level-deep references' anchor.	5 / 5
	Total	16 / 20 Passed

Description

75%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 clear, third-person description that answers both what and when with concrete triggers and a distinctive domain framing. It is strong but not top-tier because the actions stay somewhat abstract and the when-clause lacks synonyms and an explicit near-miss exclusion.

Suggestions

Sharpen the action verbs beyond 'design, audit, validate' to concrete capabilities (e.g., 'create and split skills, fix trigger routing, prune dead references').

Add a 'Do not use for...' clause to exclude near misses and push distinctiveness toward a 5.

Broaden trigger coverage with synonyms users might say (e.g., 'skill authoring', 'agent skill setup', 'AGENTS/CLAUDE guidance').

Dimension	Reasoning	Score
Specificity	"Design, audit, and validate" names three concrete actions over the named domain of repository guidance across AGENTS.md, CLAUDE.md, and .agents/skills, matching the 'several specific actions; minor gaps' anchor; not 5 because the verbs are abstract rather than distinct granular actions.	4 / 5
Completeness	It states both what ("Design, audit, and validate repository guidance...") and when ("Use when creating skills, tuning descriptions, or reviewing instruction sprawl and routing"), matching the 'both present, when could be more explicit' anchor; not 5 because the when is a broad comma list without a near-miss exclusion.	4 / 5
Trigger Term Quality	"creating skills, tuning descriptions, or reviewing instruction sprawl and routing" are natural phrases users would say, giving good keyword coverage; not 5 because synonyms and extension-style terms are sparse and "instruction sprawl" is niche vocabulary.	4 / 5
Distinctiveness Conflict Risk	The AGENTS.md/CLAUDE.md/.agents/skills framing carves a clear niche with mostly distinct triggers; not 5 because broad review phrases like "reviewing instruction sprawl" could overlap with adjacent writing/editing skills.	4 / 5
	Total	16 / 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
relative_links	Relative link issues: 1 suspicious	Warning
referenced_paths_exist	Referenced path issues: 5 missing	Warning

	Total	14 / 16 Passed

Repository: EpicenterHQ/epicenter
Path: .agents/skills/agent-instructions/SKILL.md
Commit: 9e1f9d2

Reviewed: 1 day 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.