improve-codebase-architecture

Scan a codebase for deepening opportunities, present them as a visual HTML report, then grill through whichever one you pick.

Quality

—

Does it follow best practices?

Impact

—

No eval scenarios have been run

Securityby

Advisory

Suggest reviewing before use

Quality

Content

65%

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

The content is concise, well-sequenced, and largely actionable, but its two weakest points are the missing HTML-REPORT.md bundle file (which the body relies on for the report scaffold) and the absence of validation steps around inline side effects. Adding the referenced file and a verification checkpoint would lift the lower dimensions.

Suggestions

Provide the referenced HTML-REPORT.md bundle file (or inline the report scaffold) so the 'See [HTML-REPORT.md]' link resolves.

Add an explicit verification step after writing the HTML report and after each CONTEXT.md mutation (e.g. confirm the file opened, re-read CONTEXT.md to confirm the term landed) to close the workflow_clarity validation gap.

Add a short trigger phrase in the body echoing natural user language to complement the jargon-heavy description.

Dimension	Reasoning	Score
Conciseness	The body is lean and directive throughout — it points to the /codebase-design vocabulary rather than re-explaining it, and avoids explaining concepts Claude already knows; every section earns its place. Not 2 because there is no padding or restatement of basic programming concepts.	3 / 3
Actionability	Gives concrete, executable guidance (Agent tool with subagent_type=Explore, specific temp-dir path pattern, xdg-open/open/start commands, Tailwind/Mermaid via CDN, explicit card fields), but the core deliverable scaffold is deferred to 'See [HTML-REPORT.md](HTML-REPORT.md)' — a referenced file that is not present in the bundle. Not 3 because the key HTML scaffold details are missing rather than inline.	2 / 3
Workflow Clarity	The Explore → Present → Grilling loop is clearly sequenced with a gating checkpoint ('Do NOT propose interfaces yet. After the file is written, ask the user'), but there are no validation/verification steps for the inline side effects (writing the HTML report, modifying CONTEXT.md), which the guidelines say caps workflow clarity at 2. Not 1 because the sequence and checkpoints are present.	2 / 3
Progressive Disclosure	The body is well-organized into clear sections and signals one-level-deep navigation to HTML-REPORT.md, but that referenced file does not exist in the bundle, so the disclosed reference is broken. Not 3 because the single external reference target is absent; not 1 because the main file itself is well structured.	2 / 3
	Total	9 / 12 Passed

Description

67%

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

The description is specific and distinctive in its three-phase shape, but it leans on project jargon ('deepening opportunities', 'grill through') and omits any explicit 'Use when' trigger guidance. Adding natural trigger terms and a when-to-use clause would raise the weaker dimensions.

Suggestions

Add an explicit 'Use when...' clause naming natural triggers (e.g. 'Use when reviewing architecture, finding refactoring opportunities, or surfacing design friction').

Replace or supplement jargon like 'deepening opportunities' and 'grill through' with terms users actually say, such as 'refactoring opportunities' and 'walk through design tradeoffs'.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions — 'Scan a codebase for deepening opportunities', 'present them as a visual HTML report', 'then grill through whichever one you pick' — matching the anchor for enumerating several specific actions.	3 / 3
Completeness	Clearly states what the skill does, but there is no 'Use when...' clause or equivalent explicit trigger guidance, which caps completeness at 2 per the judging guidelines.	2 / 3
Trigger Term Quality	Contains some relevant keywords ('codebase', 'HTML report') but relies on jargon like 'deepening opportunities' and 'grill through' rather than natural terms a user would say ('review architecture', 'find refactors').	2 / 3
Distinctiveness Conflict Risk	The three-part scan → visual HTML report → grilling loop is a distinctive niche unlikely to trigger for the wrong skill; not above because it could still overlap loosely with general architecture/code-review skills.	3 / 3
	Total	10 / 12 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 missing	Warning

	Total	15 / 16 Passed

Repository: coder/agent-tty
Commit: fae02cb

Reviewed: 3 days 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.