Content
73%Weight 40%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
A well-sequenced, highly actionable read-only retrieval protocol with strong workflow clarity and good external references. Its main weakness is conciseness — the QMD and GraphRAG sections are verbose and could be trimmed or split into reference files.
Suggestions
Tighten the QMD semantic-pass section: collapse the three CLI mode variants (quality/balanced/fast) into a compact table and move the extended caveats (operator-like token handling, _raw/ defensive filter) into a reference file.
Move the GraphRAG output-field reference and the structural-intents routing table into a separate reference doc, leaving SKILL.md a concise overview that links out for the field-by-field detail.
Trim the opening framing ('The wiki contains pre-synthesized, cross-referenced knowledge') and other mildly over-explanatory asides that restate what the wiki is rather than what to do with it.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body assumes Claude's competence (no 'what is a wiki/grep' explanations) and is operationally dense, but the QMD section (three CLI variants plus caveats) and the GraphRAG output-field reference could be tightened, matching the 'mostly efficient but could be tightened' 3 anchor rather than the lean 4. | 3 / 5 |
Actionability | Provides concrete executable commands throughout — 'obsidian-wiki graph-query "$OBSIDIAN_VAULT_PATH" "<question>" --pretty', 'Grep -A 10 -B 2', QMD CLI invocations, and a precise log.md format — with only minor gaps from template placeholders and conditional transport-selection logic, fitting the 4 anchor. | 4 / 5 |
Workflow Clarity | A clearly sequenced numbered pipeline (Step 0 through Step 6) with explicit decision trees, feedback loops (fallback when obsidian-wiki is absent, frontier caps, escalation reporting), and verification via staleness/trust annotations and retrieval accounting — matching the 5 anchor's explicit checkpoints and error-recovery loops; the destructive/batch cap does not apply since the skill is read-only. | 5 / 5 |
Progressive Disclosure | Good section structure with well-signaled references to shared protocols ('follow the Config Resolution Protocol in llm-wiki/SKILL.md', the local $qmd-cli skill, .env.example); no bundle files are provided and some reference-like material (QMD CLI variants, GraphRAG output fields) is inlined, which keeps it just below the 5 anchor. | 4 / 5 |
Total | 16 / 20 Passed |