CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

search

Answer questions from the user's own files and chat or mail archives through the AutoRAG tools. Use when the user asks to find, summarize, compare or verify something in their documents (PDF, DOCX, PPTX, XLSX, HWP/HWPX, Markdown, email) or in connected Slack, Discord, Telegram, WhatsApp, KakaoTalk or email archives. Not for web search.

Quality

98%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Low

Low-risk findings worth noting

SKILL.md
Quality
Evals
Security

Quality

Content

100%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 exemplary instruction-only skill body: it assumes Claude's competence, gives fully concrete tool calls and parameters, sequences the work with explicit error-recovery checkpoints, and stays within a single well-organized file that matches its narrow scope. No weaknesses found against any anchor.

DimensionReasoningScore

Conciseness

The ~49-line body is lean throughout: it never explains concepts Claude already knows (no 'what is a PDF' padding) and every sentence carries tool-specific guidance such as 'retrieval is lexical plus semantic, not reasoning' and 'topK (8 is a good start)'. Each token earns its place; nothing could be cut without losing actionable information.

5 / 5

Actionability

Fully executable instruction-only guidance: exact tool names (`autorag.search`, `autorag.datasources.list`, `autorag.search.files`, `autorag.evidence`), concrete parameters (`query`, `topK`, `datasourceIds`, `local: true`, `sessionId`, `strict: true`), and specific recipes like 'search again with two or three differently worded queries' and routing rules per question type. Per the rubric's scoring notes, absence of code in an instruction-only skill is not penalized when the guidance is this concrete.

5 / 5

Workflow Clarity

A clearly sequenced six-step workflow with explicit validation checkpoints and feedback loops: 'If a search returns `index-not-ready`, tell the user... and stop', 'If the answer is not there, search again with two or three differently worded queries', and a dedicated 'Reading the result' section mapping each diagnostic (stale, unsearched, datasource-*, embedding-unavailable) to a specific response. The skill is read-only, so the destructive/batch cap does not apply.

5 / 5

Progressive Disclosure

A single-file skill under 50 lines with no bundle files (references/, scripts/, assets/ are absent) and no need for external references; per the rubric guideline, this can score 5 with well-organized sections. The intro, Workflow, Reading the result, and Safety sections are clearly delimited and easy to navigate.

5 / 5

Total

20

/

20

Passed

Description

95%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 nails triggers, completeness, and distinctiveness: it states what the skill does, gives explicit 'Use when...' guidance with natural phrasing and a full format/platform inventory, and draws a sharp boundary against web search. The only minor gap is that the skill's own actions are summarized in a single verb ('answer questions') rather than enumerated, which leaves specificity just short of the top anchor.

DimensionReasoningScore

Specificity

Concrete actions are present — 'find, summarize, compare or verify something' plus an enumerated format and archive list (PDF, DOCX, PPTX, XLSX, HWP/HWPX, Slack, Discord, ...) — but the skill's own action is stated only once ('Answer questions... through the AutoRAG tools'), so coverage sits just below the comprehensive anchor. It is noticeably above the score-3 anchor ('names domain and 1-2 concrete actions') but does not list multiple distinct skill-side operations like the score-5 example.

4 / 5

Completeness

Explicitly answers both questions: 'what' is 'Answer questions from the user's own files and chat or mail archives through the AutoRAG tools' and 'when' is 'Use when the user asks to find, summarize, compare or verify something in their documents... or in connected... archives'. Both halves come with concrete trigger phrases, matching the score-5 anchor exactly; the score-4 anchor's 'when could be more explicit' does not apply.

5 / 5

Trigger Term Quality

Comprehensive natural-language coverage including synonyms and extensions: users would literally say 'find', 'summarize', 'compare', 'verify', 'documents', 'email', and name the platforms (Slack, Discord, Telegram, WhatsApp, KakaoTalk) and formats (PDF, DOCX, PPTX, XLSX, Markdown). This matches the score-5 anchor ('PDF files, PDFs, forms, document extraction, .pdf') in breadth.

5 / 5

Distinctiveness Conflict Risk

A clear niche (the user's own files and connected chat/mail archives via AutoRAG) with distinct triggers and an explicit exclusion — 'Not for web search' — that removes the most likely conflict. Platform and format names further separate it from generic document skills; this fits the score-5 'clear niche with distinct triggers; minimal conflict risk' anchor.

5 / 5

Total

19

/

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

plugins/remote-agents/skills/search/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.