CtrlK
BlogDocsLog inGet started
Tessl Logo

nature-polishing

Polish, translate, or tighten existing academic prose while preserving facts, terminology, and evidence boundaries. Use for 论文润色、学术翻译、正文精简, or manuscript LaTeX layout fixes. Use nature-writing when the main task is drafting new sections or rebuilding the manuscript argument.

70

Quality

88%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Passed

No findings from the security scan

SKILL.md
Quality
Evals
Security

Quality

Content

81%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.

A well-structured router skill with an explicit five-step workflow, decision checkpoints, and an exemplary lazy-loading design. The main weaknesses are editorial caveats that add tokens without guidance, and a bundle/packageging gap: the manifest.yaml that step 1 depends on is missing, which both weakens actionability and leaves most bundled reference files unreachable per the written instructions.

Suggestions

Ship manifest.yaml (and the core/ and nature-shared files it maps to) in the bundle, or inline the axis→file mapping in SKILL.md, so the routing steps are executable against the actual files.

Trim provenance caveats ("initially distilled from published NMI papers; treat them as corpus-derived guidance") into a single short clause or move them into the referenced files.

Name the on-demand references/ files (or their trigger conditions) somewhere visible when the manifest is absent, so the deep references bundled in references/ are discoverable.

DimensionReasoningScore

Conciseness

The router is lean — no explanation of what LaTeX or academic writing is, every step carries load (axes, defaults, file paths). Minor trimmable padding remains: the caveat "These defaults were initially distilled from published NMI papers; treat them as corpus-derived guidance, not official policy" and the long preserve-list sentence in step 4 are editorial caveats Claude could mostly do without. Efficient with minor over-explanation fits anchor 4, below the every-token-earns-its-place anchor 5.

4 / 5

Actionability

Highly actionable for an instruction-only skill: exact file paths (manifest.yaml, references/latex-layout.md, ../nature-shared/core/*), per-axis defaults ("Default: research", "Default: generic"), a priority order for applying fragments, and the routing rule for layout requests. Falls short of anchor 5 because the primary navigation artifact (manifest.yaml) is not present in the bundle, so steps 1–3 cannot be executed as written against the actual files.

4 / 5

Workflow Clarity

A clear five-step sequence with explicit checkpoints: "State the detected axis values in one short line to the user before proceeding, so they can correct you cheaply" (validation of routing decisions), "Ask the user if it is ambiguous and matters", and for layout jobs "Always compile and visually inspect rendered pages before and after — never judge layout from the .tex alone" (a before/after verification loop). These are explicit validation steps with correction feedback, matching anchor 5.

5 / 5

Progressive Disclosure

The design is genuinely progressive — load a manifest and small core layer, load only axis-selected fragments, "Do not read every fragment in static/", on-demand references — and references/latex-layout.md exists and is self-contained as claimed, one level deep. However, the bundle contains only references/; the manifest.yaml, core/failure-modes.md, and ../nature-shared/* files the router points to are absent, and six of the seven bundled reference files are never named in SKILL.md (they are presumably enumerated in the missing manifest). Good structure with real navigation gaps fits anchor 4 rather than the fully navigable anchor 5.

4 / 5

Total

17

/

20

Passed

Description

92%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.

A strong description: concrete actions, explicit "Use for" triggers in both Chinese and English, and an explicit disambiguation clause against the sibling nature-writing skill. The only gap is the absence of common English-language trigger synonyms beyond the Chinese phrases and the LaTeX term.

Suggestions

Add one or two natural English trigger phrases (e.g. "paper polishing", "academic proofreading") alongside the Chinese terms to catch English-only requests.

Consider naming the sibling skill family relationship once ("the drafting counterpart to this skill") is already present — no further change needed if the English triggers are added.

DimensionReasoningScore

Specificity

Lists four concrete actions — "Polish, translate, or tighten existing academic prose" plus "manuscript LaTeX layout fixes" — and adds the specific constraint "preserving facts, terminology, and evidence boundaries". This matches the anchor for multiple specific concrete actions with comprehensive coverage of the skill's niche.

5 / 5

Completeness

Explicitly answers both: what ("Polish, translate, or tighten existing academic prose while preserving facts, terminology, and evidence boundaries") and when ("Use for 论文润色、学术翻译、正文精简, or manuscript LaTeX layout fixes") with concrete trigger phrases. This mirrors the anchor-5 example structure exactly.

5 / 5

Trigger Term Quality

Triggers "论文润色、学术翻译、正文精简" and "manuscript LaTeX layout fixes" are natural, high-precision phrases for the bilingual audience, but common English variants users would actually say ("polish my paper", "academic editing", "proofread my manuscript") are absent. Good coverage with a few natural terms missing fits anchor 4, not the synonym-complete anchor 5.

4 / 5

Distinctiveness Conflict Risk

Bilingual domain-specific triggers and the explicit boundary clause "Use nature-writing when the main task is drafting new sections or rebuilding the manuscript argument" carve out a clear niche and pre-empt the one obvious sibling-skill conflict. Minimal overlap risk, matching anchor 5.

5 / 5

Total

19

/

20

Passed

Validation

93%

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

CriteriaDescriptionResult

relative_links

Relative link issues: 1 missing

Warning

Total

15

/

16

Passed

Repository
Yuan1z0825/nature-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.