research-wiki

Persistent research knowledge base that accumulates papers, ideas, experiments, claims, and their relationships across the entire research lifecycle. Inspired by Karpathy's LLM Wiki pattern. Use when user says "知识库", "research wiki", "add paper", "wiki query", "查知识库", or wants to build/query a persistent field map.

Quality

73%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./skills/research-wiki/SKILL.md

Quality

Content

63%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 thorough, largely executable skill body with strong workflow gating and concrete commands. Its main weaknesses are verbose anecdotal padding and a monolithic single-file structure that would benefit from splitting detailed reference material into bundle files.

Suggestions

Trim the historical-drift and war-story paragraphs (e.g. the 'empty for a week' anecdote) to the operative rule they illustrate.

Move large reference blocks (paper/idea/claim page schemas, the query_pack budget table, the Hook 1-4 integration pseudocode) into files under references/ and link to them, improving progressive disclosure.

Convert the Hook integration blocks from pseudocode into concrete, copy-pasteable command sequences where possible, or explicitly justify the pseudocode form.

Dimension	Reasoning	Score
Conciseness	Mostly efficient operational guidance (commands, schemas, tables), but interspersed with unnecessary narrative padding such as the historical drift anecdote about the 'prose-only init' and the war story about a user's wiki being 'empty for a week'.	3 / 5
Actionability	Subcommand sections give copy-paste-ready bash with real flags (e.g. ingest_paper --arxiv-id --thesis; add_edge --from --to --type --evidence) and a concrete page schema; the integration Hook blocks are pseudocode-style logic rather than directly executable commands, a minor gap.	4 / 5
Workflow Clarity	Subcommands are clearly sequenced with explicit validation/gating (EXP_NODE_OK gating of edges, warn-and-skip resolution, lint health checks, helper-resolution chain with fix instructions); a couple of feedback loops (lint → fix) are implicit rather than fully spelled out.	4 / 5
Progressive Disclosure	Well-organized into clear sections with clearly signaled external references (shared-references/*.md links), but the ~462-line body is monolithic — detailed reference material like the query_pack budget table, paper page schema, and Hook integration code could live in separate bundle files; no bundle files exist.	3 / 5
	Total	14 / 20 Passed

Description

83%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 that explicitly states both capabilities and trigger conditions with bilingual natural-language phrases. It is specific and distinctive; the main weakness is that the concrete subcommand actions (init/ingest/sync/update/lint/stats) are not enumerated in the description itself.

Suggestions

Consider listing one or two core subcommand verbs (e.g. init/ingest/query) in the description so the action surface is as concrete as the entity list.

Add a couple more natural English trigger phrases (e.g. 'research knowledge base', 'track papers') to broaden keyword coverage beyond the current set.

Dimension	Reasoning	Score
Specificity	Names the domain ('research knowledge base') and several concrete objects it manages ('papers, ideas, experiments, claims, and their relationships') plus explicit actions ('build/query a persistent field map'); minor gaps in action coverage since the init/ingest/sync/lint subcommands are not surfaced here.	4 / 5
Completeness	Clearly answers both 'what' ('Persistent research knowledge base that accumulates papers, ideas, experiments, claims, and their relationships') and 'when' with an explicit 'Use when user says ...' clause listing concrete trigger phrases.	5 / 5
Trigger Term Quality	Good coverage of natural trigger phrases in two languages ('知识库', '查知识库', 'research wiki', 'add paper', 'wiki query') with synonym variants; a few natural terms (e.g. 'literature', 'papers' alone) are absent, and there are no file-extension triggers.	4 / 5
Distinctiveness Conflict Risk	The wiki niche and its trigger terms are distinctive; minor overlap risk remains with closely related paper-reading/ideation skills that also touch the wiki as a side effect.	4 / 5
	Total	17 / 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
allowed_tools_field	'allowed-tools' contains unusual tool name(s)	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning
relative_links	Relative link issues: 3 suspicious	Warning

	Total	13 / 16 Passed

Repository: wanshuiyin/Auto-claude-code-research-in-sleep
Path: skills/research-wiki/SKILL.md
Commit: f4f20f9

Reviewed: 1 day 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.