CtrlK
BlogDocsLog inGet started
Tessl Logo

instruction-guardian

Use before editing CLAUDE.md, AGENTS.md, MEMORY.md, memory topic files, or any file under `.claude/rules/` — at any path depth, including nested subdirectory instances. Also when the user says "add this to CLAUDE.md", "remember this", "note this in the rules", or similar. Triggers on any change regardless of size and must not be skipped during multi-file edit sessions.

77

1.73x
Quality

74%

Does it follow best practices?

Impact

66%

1.73x

Average score across 3 eval scenarios

SecuritybySnyk

Passed

No known issues

Fix and improve this skill with Tessl

tessl review fix ./instruction-guardian/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

77%

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 guardian/gatekeeper skill with excellent actionability and workflow clarity — the 6-step checklist, decision flowchart, condensation examples, and anti-pattern table give Claude precise, executable guidance. Its main weaknesses are moderate verbosity (the rationalizations table and cleanup exception section are extensive) and the monolithic structure that could benefit from splitting detailed reference material into separate files. The skill effectively teaches a non-obvious, high-value behavior pattern that Claude would not know by default.

Suggestions

Extract the rationalizations table and the instruction-cleanup exception details into a separate reference file (e.g., EDGE_CASES.md) and reference it with a pitch-style pointer — these are consulted rarely but add significant length.

Consider moving the condensation patterns and anti-patterns tables into a companion PATTERNS.md file, keeping only 1-2 inline examples in the main skill body to illustrate the principle.

DimensionReasoningScore

Conciseness

The skill is thorough and most content earns its place, but there is notable verbosity: the rationalizations table is extensive (10 rows covering edge cases Claude could infer), the exception/carve-out section for instruction-cleanup is very detailed with multiple bullet points about flag files, and some explanations are repetitive (e.g., the spirit-vs-letter point is made multiple times). However, the condensation patterns and anti-patterns tables are genuinely useful reference material.

2 / 3

Actionability

The skill provides a concrete 6-step checklist with specific thresholds (200 lines), exact format rules per destination, copy-paste-ready condensation examples (BAD vs GOOD), a decision flowchart in DOT notation, and a specific pushback script template. Every section gives Claude precise, executable guidance rather than vague direction.

3 / 3

Workflow Clarity

The 6-question pre-write checklist is clearly sequenced with explicit ordering ('answer these six questions in order'), includes validation gates (line count check, duplication check, litmus test), has a decision flowchart for routing, and provides clear feedback loops (if over threshold → extract first, if user insists → comply but condense). The workflow handles error/edge cases explicitly.

3 / 3

Progressive Disclosure

The skill is a single monolithic file at ~200+ lines with no references to supporting documents. The rationalizations table, condensation patterns, anti-patterns table, and the detailed instruction-cleanup exception could be split into separate referenced files. However, the internal structure with clear headers and sections partially compensates.

2 / 3

Total

10

/

12

Passed

Description

72%

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 excels at defining when the skill should be triggered, with excellent trigger terms and distinctive file-based scope. However, it critically lacks any explanation of what the skill actually does — it reads entirely as a 'when to use' clause without a 'what it does' clause. This makes it difficult for Claude to understand the purpose or value of invoking this skill.

Suggestions

Add a clear 'what it does' statement at the beginning describing the concrete actions performed (e.g., 'Validates formatting, enforces structure conventions, and ensures consistency when editing memory and configuration files').

Restructure to lead with capabilities before the trigger conditions — e.g., 'Enforces formatting rules and structural conventions for CLAUDE.md, AGENTS.md, MEMORY.md, and .claude/rules/ files. Use before editing...'

DimensionReasoningScore

Specificity

The description names specific files (CLAUDE.md, AGENTS.md, MEMORY.md, memory topic files, .claude/rules/) and the domain of editing configuration/memory files, but it describes *when* to use the skill rather than *what concrete actions* the skill performs. It lacks explicit capability verbs beyond 'editing'.

2 / 3

Completeness

The 'when' is thoroughly covered with explicit triggers and file patterns, but the 'what does this do' is essentially missing — the description never explains what the skill actually does when triggered (e.g., validates formatting, enforces structure, applies conventions). It only describes when to invoke it.

2 / 3

Trigger Term Quality

Includes strong natural trigger terms users would actually say: 'add this to CLAUDE.md', 'remember this', 'note this in the rules', plus specific file names and paths like CLAUDE.md, AGENTS.md, MEMORY.md, .claude/rules/. Good coverage of natural language variations.

3 / 3

Distinctiveness Conflict Risk

Highly distinctive with very specific file targets (CLAUDE.md, AGENTS.md, MEMORY.md, .claude/rules/) that form a clear niche around Claude configuration and memory files. Unlikely to conflict with other skills due to the precise file-based triggers.

3 / 3

Total

10

/

12

Passed

Validation

100%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation11 / 11 Passed

Validation for skill structure

No warnings or errors.

Repository
ivan-magda/instruction-health-skills
Reviewed

Table of Contents

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.