mcollina/init
Creates, updates, or optimizes an AGENTS.md file for a repository with minimal, high-signal instructions covering non-discoverable coding conventions, tooling quirks, workflow preferences, and project-specific rules that agents cannot infer from reading the codebase. Use when setting up agent instructions or Claude configuration for a new repository, when an existing AGENTS.md is too long, generic, or stale, when agents repeatedly make avoidable mistakes, or when repository workflows have changed and the agent configuration needs pruning. Applies a discoverability filter—omitting anything Claude can learn from README, code, config, or directory structure—and a quality gate to verify each line remains accurate and operationally significant.
85
94%
Does it follow best practices?
Impact
72%
1.14xAverage score across 5 eval scenarios
Passed
No known issues
Quality
Discovery
100%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
This is an excellent skill description that thoroughly covers what the skill does, when to use it, and what makes it unique. It uses third person voice consistently, provides multiple explicit trigger scenarios, and clearly differentiates itself through the specific AGENTS.md focus and discoverability filter concept. The description is detailed without being padded, with every clause adding meaningful information.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Lists multiple specific concrete actions: 'Creates, updates, or optimizes an AGENTS.md file', 'applies a discoverability filter', 'quality gate to verify each line remains accurate'. Also specifies what content it covers: 'non-discoverable coding conventions, tooling quirks, workflow preferences, and project-specific rules'. | 3 / 3 |
Completeness | Clearly answers both 'what' (creates/updates/optimizes AGENTS.md with minimal high-signal instructions) and 'when' with explicit triggers ('Use when setting up agent instructions...when an existing AGENTS.md is too long...when agents repeatedly make avoidable mistakes...when repository workflows have changed'). | 3 / 3 |
Trigger Term Quality | Includes strong natural trigger terms users would say: 'AGENTS.md', 'agent instructions', 'Claude configuration', 'repository', 'agent configuration', 'coding conventions', 'workflow preferences'. Covers variations like 'too long, generic, or stale' and 'agents repeatedly make avoidable mistakes'. | 3 / 3 |
Distinctiveness Conflict Risk | Highly distinctive niche focused specifically on AGENTS.md files and agent configuration for repositories. The discoverability filter concept and focus on non-inferrable conventions make it clearly distinguishable from general documentation or configuration skills. | 3 / 3 |
Total | 12 / 12 Passed |
Implementation
85%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a well-crafted skill that is concise, well-structured, and provides clear decision frameworks (discoverability filter, quality gate) for creating AGENTS.md files. Its main weakness is the lack of concrete examples—a sample AGENTS.md snippet showing good vs. bad entries would significantly improve actionability. The workflow is logical and includes appropriate validation steps.
Suggestions
Add a concrete example showing 2-3 good AGENTS.md entries alongside 2-3 bad entries that fail the discoverability filter, to make the guidance more immediately executable.
Include a short template or skeleton of a well-formed AGENTS.md file that Claude can use as a starting point when creating one from scratch.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Every section earns its place. There's no explanation of what AGENTS.md is in general terms or what agents are—it jumps straight into actionable filtering criteria and structural guidance. The content is lean and assumes Claude's intelligence throughout. | 3 / 3 |
Actionability | The guidance is specific and well-structured with clear decision criteria (discoverability filter, quality gate), but it lacks concrete examples of actual AGENTS.md output. There are no before/after examples or a template showing what a good AGENTS.md looks like, which would make it more directly executable. | 2 / 3 |
Workflow Clarity | The workflow is clear: check source files first, apply the discoverability filter per-line, structure output using recommended sections, then run the quality gate before finalizing. The quality gate serves as an explicit validation checkpoint with a delete-if-fails feedback loop. For this type of non-destructive content generation task, this level of workflow clarity is sufficient. | 3 / 3 |
Progressive Disclosure | For a standalone skill with no bundle files, the content is well-organized into clearly labeled sections that progress logically from when to use, to core rules, to what to include/exclude, to structure, to source files, to maintenance, to final validation. No external references are needed for this scope. | 3 / 3 |
Total | 11 / 12 Passed |
Validation
90%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 10 / 11 Passed
Validation for skill structure
| Criteria | Description | Result |
|---|---|---|
metadata_version | 'metadata.version' is missing | Warning |
Total | 10 / 11 Passed | |
Reviewed
Table of Contents