CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

agents-llamaindex

Data framework for building LLM applications with RAG. Specializes in document ingestion (300+ connectors), indexing, and querying. Features vector indices, query engines, agents, and multi-modal support. Use for document Q&A, chatbots, knowledge retrieval, or building RAG pipelines. Best for data-centric LLM applications.

Quality

77%

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

Fix and improve this skill with Tessl

tessl review fix ./cli-tool/components/skills/ai-research/agents-llamaindex/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

67%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 highly actionable, well-structured reference whose code examples are nearly all executable and well sequenced along the RAG pipeline, with real one-level-deep reference files. Its weaknesses are token bloat from marketing metrics, benchmarks, and duplicated comparison content, and an oversized body that should offload detail to the existing bundle files.

Suggestions

Remove the metrics block (GitHub stars, contributors), the performance benchmarks table, and the marketing tagline — they consume context without guiding execution.

Trim the duplication between "When to use LlamaIndex" and the "LlamaIndex vs LangChain" section into a single routing block, and move the vector store integrations and multi-modal sections into reference files like the existing references/ guides.

Fix the non-executable snippets: complete the custom retriever example (define how nodes are produced) and update the Pinecone example to the current client API.

DimensionReasoningScore

Conciseness

The majority of the body is dense, executable code without re-explaining concepts Claude already knows, but several sections are pure padding: "The leading framework for connecting LLMs with your data", the GitHub-stars/contributor metrics block, the latency benchmark table, the Discord link, and a LlamaIndex-vs-LangChain section that duplicates the "When to use LlamaIndex" list twice. Time-sensitive figures (v0.14.7, 45,100+ stars) are not quarantined in an old-patterns section, which the guidelines say should penalize conciseness. Not 2 because most tokens are working code; not 4 because the unnecessary sections are more than minor trims.

3 / 5

Actionability

Nearly every section provides complete, copy-paste-ready code with real values covering the common cases (install, 5-line RAG, connectors, agents, chat engine, metadata filters, vector stores, structured output, evaluation). Not 5 because the custom retriever example is non-executable pseudocode ("return nodes" with undefined nodes) and the Pinecone example uses the deprecated pinecone.init() v2 client API; not 3 because those gaps are isolated in an otherwise fully executable body.

4 / 5

Workflow Clarity

Core concepts are sequenced along the natural ingest -> index -> persist -> query pipeline, the quick start is unambiguous, end-to-end patterns (Document Q&A, chatbot with memory) are complete, and the Evaluation section (RelevancyEvaluator/FaithfulnessEvaluator) provides a response-quality checkpoint. Not 5 because there are no explicit validation or error-recovery loops around fragile steps such as index persistence or storage loading.

4 / 5

Progressive Disclosure

Sections are well organized with clear headers, and a References section clearly signals three real, one-level-deep bundle files (references/query_engines.md, references/agents.md, references/data_connectors.md all exist, none nest further references). Not 5 because the ~570-line body inlines substantial detail — vector store integrations, ingestion patterns, multi-modal RAG, evaluation — that belongs in bundle files, leaving the body heavier than an overview.

4 / 5

Total

15

/

20

Passed

Description

78%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 answers both what the skill does and when to use it, with mostly natural trigger terms and a distinct RAG/data niche. Its main weakness is specificity: the listed capabilities are domain categories rather than concrete, fine-grained actions, and it includes mild marketing abstraction.

Suggestions

Replace category-level verbs with concrete actions, e.g. "Indexes documents into vector stores and queries them with retrieval-augmented generation" instead of "Specializes in document ingestion, indexing, and querying".

Drop the abstract framing ("Data framework for building LLM applications with RAG", "Best for data-centric LLM applications") in favor of adding missing natural synonyms such as "retrieval-augmented generation" or "semantic search".

DimensionReasoningScore

Specificity

"Specializes in document ingestion (300+ connectors), indexing, and querying" names several domain operations, but they are category-level verbs rather than concrete actions, and "Data framework for building LLM applications with RAG" / "Best for data-centric LLM applications" are abstract framing. It exceeds anchor 3's 1-2 concrete actions in count but lacks the concreteness of anchor 4's example actions, so it sits at 3.

3 / 5

Completeness

Both parts are explicit: the "what" ("Specializes in document ingestion, indexing, and querying. Features vector indices, query engines, agents, and multi-modal support") and a concrete "when" clause ("Use for document Q&A, chatbots, knowledge retrieval, or building RAG pipelines") listing four trigger use cases. Not 4 because the "when" is already explicit and specific, matching the anchor-5 pattern.

5 / 5

Trigger Term Quality

Natural user phrases like "RAG", "document Q&A", "chatbots", "knowledge retrieval", and "building RAG pipelines" give good keyword coverage. Not 5 because common variations such as "retrieval-augmented generation", "semantic search", and "embeddings" are absent; not 3 because the natural terms users would actually say are well represented.

4 / 5

Distinctiveness Conflict Risk

The RAG/document-ingestion/data-framework niche is fairly distinct with specific triggers (connectors, vector indices, query engines). Not 5 because broad terms like "agents", "chatbots", and "LLLM applications" create minor overlap risk with general agent-framework skills; not 3 because the core triggers are firmly outside that overlap.

4 / 5

Total

16

/

20

Passed

Validation

87%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 14 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

skill_md_line_count

SKILL.md is long (571 lines); consider splitting into references/ and linking

Warning

metadata_field

'metadata' should map string keys to string values

Warning

Total

14

/

16

Passed

Repository

davila7/claude-code-templates

Path

cli-tool/components/skills/ai-research/agents-llamaindex/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.