CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

autorag-doctor

Diagnose and repair a broken or half-working AutoRAG install so every configured source is both indexed and searchable. Checks AutoRAG, MinSync, Jikji, the embedding backend, and every CLI-backed datasource (lazykatok, discrawl, slacrawl, wacrawl, telecrawl, notcrawl, mailcrawl, rclone, Spotlight), then fixes orphan locks, orphan processes, embedding dimension or identity mismatches, stale indexes, and missing setup. Use when search returns nothing or too little, refresh hangs or fails, a datasource disappeared from results, indexes look stale, the embedder is unavailable, or the user asks to check, diagnose, verify, or repair AutoRAG.

Quality

95%

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

88%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.

An excellent operational playbook: fully executable commands for every datasource, a tightly sequenced diagnose-and-verify workflow with real validation gates, and no filler. The main weaknesses are mild redundancy between the repair playbook and the diagnostic-codes table, and a monolithic single-file structure that could offload the codes table to a reference file.

Suggestions

Consolidate the duplicated diagnostic explanations: 'jikji-unavailable', 'embedder-unavailable', and 'datasource-auth-error' are described in both the 'Missing or wrong setup' section and the diagnostic codes table — keep one canonical entry and cross-reference it.

Move the diagnostic-codes table (and optionally the per-CLI repair detail) into a one-level-deep reference file (e.g. references/diagnostic-codes.md), keeping SKILL.md as the triage-and-workflow overview with well-signaled links.

The 'docs/minsync-setup.md' pointer is not part of the skill bundle — either add it as a bundled reference or reword to point at where the user can actually find it (e.g. the AutoRAG install directory).

DimensionReasoningScore

Conciseness

Efficient and dense — nearly every sentence is an AutoRAG-specific operational fact Claude could not know, with commands, paths, and diagnostic codes rather than concept explanations. Not 5: a few rules are wordy, and 'jikji-unavailable', 'embedder-unavailable', and 'datasource-auth-error' are each explained twice (once in the repair playbook, again in the diagnostic codes table), which could be consolidated.

4 / 5

Actionability

Fully executable throughout: a per-datasource table of exact native-check and re-sync commands, copy-paste refresh/rebuild/retry commands, concrete MCP invocations with arguments ('autorag.search {"query":...,"topK":3}', 'autorag.evidence {"sessionId":...}'), and a worked SQL cross-check example. Specific examples cover the common failure cases.

5 / 5

Workflow Clarity

A clear triage → probe → prove-searchability → repair → report sequence with explicit validation checkpoints: repairs must be re-verified by re-running the failing check, 'Re-run autorag status --json afterwards to confirm inFlight: false', never kill a running sync, and a completion condition gates the run on real query hits. The destructive operations (kills, lock deletion, index reset) all carry confirm-first validation, so the destructive/batch cap does not apply.

5 / 5

Progressive Disclosure

Well-sectioned single file with no nested references and clear navigation (triage, per-datasource table, repair playbook, codes table, report). Not 5: the ~280-line body keeps the full diagnostic-codes table and repair playbook inline where well-signaled one-level references could carry them, and the lone external pointer 'docs/minsync-setup.md' is not part of the bundle. Not 3: structure is good and inline placement is defensible for a repair run that needs the codes at hand.

4 / 5

Total

18

/

20

Passed

Description

100%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 model description: concrete what, explicit when with natural trigger phrases, named components and failure classes, and a distinct niche with essentially no conflict risk. Third-person voice throughout, with no fluff or over-claims.

DimensionReasoningScore

Specificity

Lists multiple concrete actions — 'Diagnose and repair', 'Checks AutoRAG, MinSync, Jikji, the embedding backend, and every CLI-backed datasource', 'fixes orphan locks, orphan processes, embedding dimension or identity mismatches, stale indexes, and missing setup' — with comprehensive coverage of both components and failure modes. It exceeds the 'minor gaps' bar for 4 because the enumerated fixes map one-to-one to the skill's repair scope.

5 / 5

Completeness

Explicitly answers 'what' (diagnose and repair an AutoRAG install, checking named components and fixing named failure classes) and 'when' via a concrete 'Use when...' clause enumerating six trigger scenarios. This matches the 5 anchor example's structure of full what + explicit trigger phrases.

5 / 5

Trigger Term Quality

Natural user phrases are comprehensively covered: 'search returns nothing or too little', 'refresh hangs or fails', 'a datasource disappeared from results', 'indexes look stale', 'the embedder is unavailable', plus the synonym set 'check, diagnose, verify, or repair'. No common variation users would say when facing a broken AutoRAG install is missing.

5 / 5

Distinctiveness Conflict Risk

The entire description is anchored on AutoRAG and its specific backing CLIs (MinSync, Jikji, lazykatok, discrawl, etc.), a clear niche with distinct triggers; it would not plausibly fire for a non-AutoRAG skill.

5 / 5

Total

20

/

20

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 — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository

Marker-Inc-Korea/AutoRAG

Path

skills/autorag-doctor/SKILL.md

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.