instruction-guardian

Routing gate for instruction-file edits that decides where new content belongs and pushes back on bloat. Use before ANY edit, of any size, even mid multi-file sweeps, to CLAUDE.md, AGENTS.md, MEMORY.md, memory topic files, or any file under `.claude/rules/`, at any path depth including nested instances. Also when the user says "add this to CLAUDE.md", "remember this", "note this in the rules", or similar. For bulk restructuring of an oversized file, use instruction-cleanup instead.

1.12x

Quality

88%

Does it follow best practices?

Impact

99%

1.12x

Average score across 3 eval scenarios

Securityby

Passed

No findings from the security scan

Quality

Content

77%Weight 40%Scale 1-3

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 a clear sequenced checklist and validation gates, but is overly long and monolithic with no external references to offload detail. Tightening the repetitive exception prose and splitting the large tables into referenced files would improve token efficiency and disclosure.

Suggestions

Condense the repetitive exception/spirit-vs-letter prose — the carve-out is explained across several overlapping sections; consolidate to one and reference it.

Move the large 'Rationalizations to catch in yourself' and 'Anti-Patterns to Catch' tables into a reference file (e.g. references/rationalizations.md) with a one-line pitch-style pointer, improving both conciseness and progressive disclosure.

Trim redundancy between the activation section and the rationalizations table, which restate the same 'trigger is the file, not the tool / size' point multiple times.

Dimension	Reasoning	Score
Conciseness	It avoids explaining concepts Claude already knows and uses compact BAD/GOOD examples, but the ~250-line body is padded with repetitive exception mechanics and a large rationalizations table that restates the spirit-vs-letter clause several times; it could be tightened.	2 / 3
Actionability	Provides a concrete 6-question checklist, a threshold table with exact line counts, a copy-paste pushback script, 'X not Y because Z' format, a routing flowchart, and BAD/GOOD examples — fully executable guidance.	3 / 3
Workflow Clarity	A clearly sequenced 6-step pre-write checklist with explicit validation gates (over-threshold → extract/condense first) and a push-back/comply-and-condense feedback loop for routing failures.	3 / 3
Progressive Disclosure	No bundle files exist (references/scripts/assets absent) and all detail is inline in a single ~250-line SKILL.md with no one-level-deep references; sections are well-organized but content such as the rationalizations and anti-patterns tables could be split out.	2 / 3
	Total	10 / 12 Passed

Description

100%Weight 40%Scale 1-3

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, trigger-rich, and clearly answers both what and when, with strong distinctiveness from sibling skills. It is a strong, well-scoped activation description.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions ("decides where new content belongs", "pushes back on bloat") and concrete file targets (CLAUDE.md, AGENTS.md, MEMORY.md, .claude/rules/), matching the 'multiple specific concrete actions' anchor.	3 / 3
Completeness	Explicitly states what the skill does (routing gate that decides destination and pushes back on bloat) and when to use it ("Use before ANY edit", "Also when the user says...") with explicit triggers, in third person.	3 / 3
Trigger Term Quality	Includes natural phrases a user would actually say — "add this to CLAUDE.md", "remember this", "note this in the rules" — giving good coverage of real trigger terms.	3 / 3
Distinctiveness Conflict Risk	Occupies a clear niche (instruction-file edit routing) with distinct triggers and explicitly distinguishes from the sibling skill instruction-cleanup, making wrong-skill triggering unlikely.	3 / 3
	Total	12 / 12 Passed

Validation

100%

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 — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: ivan-magda/instruction-health-skills
Path: instruction-guardian/SKILL.md
Commit: 2837bdc

Reviewed: 14 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.