he-improve

Improve existing Harness Engineering implementations or workflows with evidence-backed changes. Use when users ask for targeted enhancement of shipped or drafted work.

Quality

—

Does it follow best practices?

Impact

—

No eval scenarios have been run

Securityby

Passed

No known issues

Quality

Content

62%

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

The content is well-organized with a strong, explicitly-validated workflow, but its procedural guidance stays abstract without concrete templates or commands, and its progressive-disclosure promise is undermined by references to files that are not present.

Suggestions

Add a concrete spec template or example measurement command and a sample `schema_version: 1` output object so the Procedure steps are executable rather than abstract.

Fix or remove the broken references: either ship references/session-evidence-contract.md and the assets/icon-*.png files or drop the links so navigation is not dangling.

Collapse the Anti-patterns and Constraints sections into the Procedure/Validation they restate to remove redundancy and improve token efficiency.

Dimension	Reasoning	Score
Conciseness	The body is mostly directive and avoids explaining concepts Claude already knows, but sections like Anti-patterns and Constraints substantially restate the Procedure and Validation in negative form, so it could be tightened; it does not reach the 'every token earns its place' level-3 anchor.	2 / 3
Actionability	It provides a structured procedure and validation checklist (concrete enough to be instructions), but steps like 'Load or create the optimization spec' and 'Run bounded iterations with explicit measurement gates' stay abstract with no spec template, command, or schema example, leaving key details missing as the level-2 anchor describes.	2 / 3
Workflow Clarity	The 8-step Procedure is clearly sequenced with explicit validation checkpoints (step 7's write-then-verify, the Validation 'Fail fast: stop at first failed gate') and a keep/revise/discard feedback loop, matching the level-3 anchor for clear sequence with explicit validation and error-recovery loops.	3 / 3
Progressive Disclosure	The body is well-sectioned and claims to keep context in 'archived references', but the referenced file '../../../../../../references/session-evidence-contract.md' and the Full-Context assets (icon-small.png, icon-large.png) do not exist, so the one-level-deep navigation is broken rather than clean.	2 / 3
	Total	9 / 12 Passed

Description

75%

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 cleanly answers both what and when with an explicit trigger and a distinct domain niche, but its action vocabulary is broad and its trigger phrasing is somewhat internal/jargon-heavy rather than natural user speech.

Suggestions

List two or three concrete actions (e.g., 'tune metrics, compare gated experiments, route proven changes') instead of the single verb 'improve' to reach specificity level 3.

Add natural-language trigger variants users actually say ('tune', 'optimize', 'make this workflow better', 'compare two approaches') alongside 'targeted enhancement'.

Dimension	Reasoning	Score
Specificity	Names the domain ('Harness Engineering implementations or workflows') and an action ('improve ... with evidence-backed changes'), but only offers one broad action rather than multiple specific concrete actions, so it stops short of the level-3 anchor.	2 / 3
Completeness	It explicitly states both what it does ('Improve existing ... with evidence-backed changes') and when to use it ('Use when users ask for targeted enhancement ...'), matching the level-3 anchor for answering what AND when with an explicit trigger.	3 / 3
Trigger Term Quality	The 'Use when users ask for targeted enhancement of shipped or drafted work' clause supplies relevant triggers, but the phrasing leans on internal jargon ('Harness Engineering', 'shipped or drafted work') and misses common user variations like 'tune', 'optimize', or 'make this better'.	2 / 3
Distinctiveness Conflict Risk	The 'Harness Engineering' domain qualifier plus 'evidence-backed changes' carve a clear niche unlikely to fire for unrelated improvement skills, matching the level-3 'clear niche with distinct triggers' anchor; it is well above the generic level-1 example.	3 / 3
	Total	10 / 12 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
metadata_version	'metadata.version' is missing	Warning
relative_links	Relative link issues: 2 missing, 1 suspicious	Warning

	Total	14 / 16 Passed

Repository: jscraik/Agent-Skills
Commit: 5a6027f

Reviewed: 5 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.