llm-wiki

The foundational knowledge distillation pattern for building and maintaining an AI-powered Obsidian wiki. Based on Andrej Karpathy's LLM Wiki architecture. Use this skill whenever the user wants to understand the wiki pattern, set up a new knowledge base, or needs guidance on the three-layer architecture (raw sources → wiki → schema). Also use when discussing knowledge management strategy, wiki structure decisions, or how to organize distilled knowledge. This is the "theory" skill — other skills handle specific operations (ingesting, querying, linting).

Quality

70%

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 ./.skills/llm-wiki/SKILL.md

Quality

Content

57%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 dense, information-rich schema/theory document with strong concrete templates and executable config-resolution code, but it is over-long for a SKILL.md overview, inlines material that belongs in reference files, and lacks validation checkpoints for the destructive operations it describes. Tightening and splitting into the existing references/ folder would materially improve it.

Suggestions

Move the Page Template, Paper Deep-Dive Template, confidence formula/source-quality tables, and the source-key contract into reference files under references/, leaving SKILL.md a concise overview with clearly signaled links (e.g. "See references/page-templates.md").

Add explicit validation/verification checkpoints to the Rebuild/Restore workflow description (archive-then-verify, dry-run before clear) so the destructive mode is not described without a feedback loop.

Trim redundancy such as the duplicated `CODE_UNDERSTANDING_CODEGRAPH_BIN` entry and condense the env-var catalog to the required variable plus a pointer to `.env.example`.

Dimension	Reasoning	Score
Conciseness	Most content is framework-specific and earns its place (provenance markers, confidence formula, source-key contract), but the ~670-line body has noticeable slack: a duplicated `CODE_UNDERSTANDING_CODEGRAPH_BIN` entry and several sections that over-explain. Not 4 because the padding is more than minor; not 2 because the bulk is genuinely non-obvious domain detail rather than concepts Claude already knows.	3 / 5
Actionability	Provides concrete, executable guidance: copy-paste shell functions (`obsidian_wiki_config_dir`, `find_config`), full page and paper-deep-dive templates, the confidence formula with a source-quality bucket table, and retrieval-primitive commands (`Grep -A <n> -B <n>`). Not 5 because some detail is delegated to companion skills (manifest schema, lint invocation) leaving minor gaps.	4 / 5
Workflow Clarity	The Config Resolution Protocol is a clear 0→1→2→3 sequence, but the destructive Rebuild mode (archive + clear + reprocess) is described without validation checkpoints in this skill, and most operational workflows are offloaded to companion skills. Capped at 3 per the destructive/batch-operations-without-validation rule; not 2 because the config sequence is genuinely coherent.	3 / 5
Progressive Disclosure	Section structure is good, but substantial content that could live in reference files is inlined (both page templates, the confidence formula + bucket table, the source-key contract, config-resolution shell functions, the env-var catalog), and the existing bundle files (references/karpathy-pattern.md, references/WRITING.md) are not clearly signaled or linked from the body. Not 4 because the references are not well-signaled and too much bulk is inline.	3 / 5
	Total	13 / 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, well-structured description that explicitly covers both what the skill does and when to use it, with clear boundary language against sibling operational skills. The main weakness is that the named actions are conceptual (understand, discuss, organize) rather than concrete operations.

Dimension	Reasoning	Score
Specificity	Names the domain ("AI-powered Obsidian wiki", "knowledge distillation pattern") and several specific actions ("building and maintaining", "understand the wiki pattern", "set up a new knowledge base", "organize distilled knowledge"), though the actions skew conceptual rather than operational. Not 5 because the actions lack the concrete operational specificity of the top anchor; not 3 because more than 1-2 specific actions are named.	4 / 5
Completeness	Clearly answers both what ("foundational knowledge distillation pattern for building and maintaining an AI-powered Obsidian wiki") and when ("Use this skill whenever the user wants to...", "Also use when discussing...") with concrete trigger phrases. Explicitly matches the top anchor for both what and when.	5 / 5
Trigger Term Quality	Good natural-term coverage ("wiki", "knowledge base", "knowledge management", "Obsidian wiki", "wiki structure", "distilled knowledge") with explicit "Use this skill whenever..." and "Also use when..." phrasings. Not 5 because synonyms and the file-extension-style variants that define the top anchor are not comprehensively enumerated.	4 / 5
Distinctiveness Conflict Risk	Carves a clear niche as the "theory" skill and explicitly states "other skills handle specific operations (ingesting, querying, linting)", which bounds it well. Not 5 because "set up a new knowledge base" overlaps with the wiki-setup companion skill, a minor conflict risk with a closely related skill.	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
skill_md_line_count	SKILL.md is long (678 lines); consider splitting into references/ and linking	Warning
relative_links	Relative link issues: 3 missing, 4 suspicious	Warning
referenced_paths_exist	Referenced path issues: 3 missing	Warning

	Total	13 / 16 Passed

Repository: Ar9av/obsidian-wiki
Path: .skills/llm-wiki/SKILL.md
Commit: 58aba18

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.