deepxiv
Search and progressively read open-access academic papers through DeepXiv. Use when the user wants layered paper access, section-level reading, trending papers, or DeepXiv-backed literature retrieval.
51
55%
Does it follow best practices?
Impact
—
No eval scenarios have been run
Advisory
Suggest reviewing before use
Optimize this skill with Tessl
npx tessl skill review --optimize ./skills/skills-codex/deepxiv/SKILL.mdQuality
Discovery
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 is structurally sound with a clear 'Use when' clause and a distinctive niche around DeepXiv. However, it could benefit from more concrete action verbs and broader natural trigger terms that users would actually say (e.g., 'research papers', 'arxiv', 'preprints'). The phrase 'layered paper access' is somewhat jargony and could be made more intuitive.
Suggestions
Add more natural trigger terms users would say, such as 'research papers', 'arxiv', 'preprints', 'scientific articles', or 'find papers'.
Replace vague phrases like 'layered paper access' and 'literature retrieval' with more concrete actions such as 'browse paper sections', 'retrieve abstracts', or 'search by topic/author'.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Names the domain (academic papers via DeepXiv) and some actions ('search', 'progressively read', 'section-level reading', 'trending papers'), but the actions are somewhat vague—'layered paper access' and 'literature retrieval' are not concrete operations like 'extract abstracts' or 'summarize findings'. | 2 / 3 |
Completeness | Clearly answers both 'what' (search and progressively read open-access academic papers through DeepXiv) and 'when' (explicit 'Use when' clause covering layered paper access, section-level reading, trending papers, or DeepXiv-backed literature retrieval). | 3 / 3 |
Trigger Term Quality | Includes some relevant keywords like 'academic papers', 'trending papers', 'literature retrieval', and 'DeepXiv', but misses common natural user terms like 'research papers', 'arxiv', 'scientific articles', 'preprints', 'paper search', or 'find papers'. | 2 / 3 |
Distinctiveness Conflict Risk | The mention of 'DeepXiv' as a specific platform, combined with distinctive features like 'progressive reading' and 'section-level reading', creates a clear niche that is unlikely to conflict with generic search or document skills. | 3 / 3 |
Total | 10 / 12 Passed |
Implementation
35%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The skill provides a structured workflow for progressive paper reading via DeepXiv with clear depth escalation, but suffers from verbosity—particularly the adapter resolution script, dual command listings, and the comparison table. Actionability is moderate: commands are provided but use placeholders rather than complete examples, and several steps (4, 6) are vague. Missing error handling and validation checkpoints weaken the workflow for a tool-integration skill.
Suggestions
Significantly trim the adapter resolution script—move it to a referenced file or reduce to a 2-line summary with a pointer to integration-contract.md for details.
Replace placeholder commands with at least one complete, copy-paste-ready example showing a full search-to-read flow (e.g., `/deepxiv "transformer attention" - max: 5` → actual command with real values).
Add error handling/validation: what to do if `deepxiv` returns empty results, if the SDK isn't installed, or if the adapter script fails at runtime.
Remove or collapse the comparison table and the dual (adapter + fallback) command blocks to reduce token overhead—a single block with inline fallback notation would suffice.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is quite verbose with unnecessary elements: the comparison table explaining other skills, the lengthy adapter resolution bash script, redundant fallback command listings, and explanatory prose that Claude doesn't need. The Constants section with policy references and the detailed chain resolution add significant token overhead. | 1 / 3 |
Actionability | The skill provides concrete bash commands for both the adapter and fallback paths, which is good. However, the commands use placeholder variables (QUERY, MAX_RESULTS, ARXIV_ID) without showing complete executable examples, and Step 4 ('present a compact literature table') and Step 6 are vague rather than concrete. | 2 / 3 |
Workflow Clarity | The 6-step workflow is clearly sequenced and the progressive depth escalation (search → brief → head → section) is well-defined. However, there are no validation checkpoints—no error handling if commands fail, no verification that results were returned, and no feedback loop for when the adapter or CLI produces errors. | 2 / 3 |
Progressive Disclosure | The skill references external files like `shared-references/integration-contract.md` appropriately, but no bundle files are provided to support these references. The main content itself is monolithic—the lengthy adapter resolution script and dual command listings (adapter + fallback) could be split into a separate reference file, keeping the SKILL.md leaner. | 2 / 3 |
Total | 7 / 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.
Validation — 11 / 11 Passed
Validation for skill structure
No warnings or errors.
- Commit
6f1302d
- 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.