dream-loop

Build stunning 3D scenes via a concept-art fidelity loop.

Quality

71%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./optional-skills/creative/dream-loop/SKILL.md

Quality

Content

88%Weight 40%Scale 1-5

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

The body is a strong, actionable process spec: a clearly sequenced 8-stage loop with concrete tool calls, a copy-paste judge prompt, explicit self-check validation, and well-defined exit criteria. Its main weakness is progressive disclosure — all content lives inline in SKILL.md with no reference files, though the structure is otherwise excellent.

Dimension	Reasoning	Score
Conciseness	The body is detailed but mostly purposeful — concrete tool calls, a full judge prompt, and specific commands assume Claude's intelligence rather than re-explaining basics. Minor florid passages (e.g. "beautiful surfaces and materials that shaders render well", the concept-art material list) could be trimmed. Not a 5 because of those few over-explained passages, not a 3 because the bulk is efficient and earns its tokens.	4 / 5
Actionability	Provides fully concrete, executable guidance throughout — exact tool sequences (image_generate, browser_exec, new_tab/wait_for_load/capture_screenshot, vision_analyze, delegate_task), real commands (`python3 -m http.server`, ImageMagick `convert concept.png shot.png +append compare.png`), and a copy-paste-ready judge prompt with a full scoring ladder. Not a 4 because the common cases are covered with specific, runnable commands rather than vague hints.	5 / 5
Workflow Clarity	A clear 8-stage sequence with an explicit self-check validation gate before judging ("only submit if confident you've significantly improved the score"), feedback loops (re-judge after optimizing for FPS, stall detection triggering a structural change), and concrete exit criteria. Fits the anchor 'clear sequence with explicit validation steps; feedback loops for error recovery'.	5 / 5
Progressive Disclosure	Well-organized with a Quick Reference table, numbered procedure sections, Pitfalls, and Verification; the long inline judge ladder is justified because it runs every round. No bundle files exist in references/scripts/assets, and the ~50-line judge prompt could arguably live in a reference file, which keeps it from a 5. Not a 3 because structure and navigation are genuinely good, not buried.	4 / 5
	Total	18 / 20 Passed

Description

53%Weight 40%Scale 1-5

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 concise and names a clear, distinctive niche, but it omits any explicit 'when to use' trigger guidance and leans on jargon ('concept-art fidelity loop') plus marketing language ('stunning') rather than natural user phrasing. Adding a 'Use when...' clause with concrete trigger terms would lift the capped completeness and trigger-term dimensions.

Suggestions

Append a 'Use when...' clause with natural triggers, e.g. 'Use when the user asks for a high-fidelity 3D game, scene, or WebGL/three.js visual built to AAA quality.'

Replace jargon/fluff ('concept-art fidelity loop', 'stunning') with concrete user-facing terms like '3D scene', 'game', 'WebGL', 'three.js'.

List 1-2 more concrete actions (e.g. 'generate concept art, build the scene, and iterate via a scored judge loop') to move specificity from 3 toward 4-5.

Dimension	Reasoning	Score
Specificity	Quotes "Build stunning 3D scenes" and "via a concept-art fidelity loop" — it names the domain (3D scenes) and one concrete action (build) plus the mechanism, but is not comprehensive; fits the anchor 'names domain and 1-2 concrete actions'. Not a 4 because only one real action verb is present, not a 2 because the domain and a concrete action are clearly named.	3 / 5
Completeness	Has a clear 'what' ("Build stunning 3D scenes via a concept-art fidelity loop") but no 'Use when...' clause or equivalent explicit trigger guidance, so per the rubric guideline completeness is capped at 3. Not a 4 because 'when' is entirely missing rather than weakly implied.	3 / 5
Trigger Term Quality	"3D scenes" is a natural term, but "concept-art fidelity loop" is jargon a user would rarely say and "stunning" is marketing fluff; missing common natural variations like 'game', 'WebGL', or 'three.js' that the body actually covers. Not a 4 because several natural user phrases are absent, not a 2 because at least one relevant keyword is present.	3 / 5
Distinctiveness Conflict Risk	"3D scenes via a concept-art fidelity loop" carves a fairly distinct niche unlikely to trigger for unrelated skills, with only minor overlap risk against general 3D/modeling skills. Not a 5 because the phrasing could still overlap with broader 3D-asset skills, not a 3 because the fidelity-loop framing is genuinely specific.	4 / 5
	Total	13 / 20 Passed

Validation

81%

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 — 13 / 16 Passed

Validation for skill structure

Criteria	Description	Result
metadata_version	'metadata.version' is missing	Warning
metadata_field	'metadata' should map string keys to string values	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	13 / 16 Passed

Repository: NousResearch/hermes-agent
Path: optional-skills/creative/dream-loop/SKILL.md
Commit: 68fa7e9

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.