More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Structured Knowledge and Complex Documents

    1

    Process difficult documents, use graphs and query tables in natural language. 10 ordered skills, starting with GraphRAG and Knowledge Graphs, Knowledge Graph Construction, Ontologies and Taxonomies.

    Free

    Security scanned
    Structured Knowledge and Complex Documents

    Structured Knowledge and Complex Documents

    Example session with this skill installed

    Fictional company "Northwind Contracts Ltd" has about 4,000 supplier contracts as text. The legal-ops team's plain vector RAG answers clause lookups well but fails on questions like "Which indemnity terms recur across suppliers in the logistics category?" and "Which suppliers share a subcontractor?".

    • Read your context and instructions
    • Compiled the structured knowledge

    GraphRAG Assessment: Northwind Contracts (fictional)

    Recommendation

    Adopt partially. Keep hybrid vector search for lookup questions (about 58% of the sample); add graph retrieval only for multi-hop and thematic questions, which plain RAG cannot answer from top-k chunks.

    Question analysis

    | Class | Count | Retrieval mode |
    | Lookup | 14 | Hybrid vector + keyword (existing) |
    | Multi-hop | 6 | Local entity-neighbor search with clause citations |
    | Thematic | 4 | Community-summary search, map then reduce |

    Graph design

    Entity types: Supplier, Subcontractor, Contract, Clause, ClauseType, Category. Relationships: HAS_CONTRACT, CONTAINS_CLAUSE, OF_TYPE, USES_SUBCONTRACTOR, IN_CATEGORY. Provenance: every node and edge stores contract ID and clause number. Resolution: strip legal suffixes, exact merge, fuzzy candidates above a reviewed threshold go to a human queue.

    Pipeline plan

    Units by clause; extraction per unit; merge; hierarchical clustering; community summaries. Cost = contracts x chunks x extraction calls + communities x summary calls (to be filled with real counts before approval). Run inside the company tenant.

    Evaluation plan

    24 tuning questions plus 10 held-out; baseline = current RAG; metrics: answer correctness, citation validity, thematic coverage; grader = two legal-ops reviewers; pass = better on multi-hop and thematic, no loss on lookup.

    Risks and open decisions

    Connects securely to your tools. The creator never sees your data.

    About this skill

    The problem

    Projects with PDFs, tables, graphs and semi-structured data often need a repeatable process, not a one-off prompt. This pack targets: Process difficult documents, use graphs and query tables in natural language.

    What's inside (10 skills, suggested learning order)

    1. GraphRAG and Knowledge Graphs — Use graphs for contextual retrieval.
    2. Knowledge Graph Construction — Model the entities and relationships of a domain.
    3. Ontologies and Taxonomies — Structure controlled vocabulary for AI.
    4. Knowledge Representation — Encode domain knowledge for AI systems.
    5. PDF Ingestion and OCR for RAG — Convert scanned documents into usable text.
    6. Complex Document Parsing — Extract tables and layout from documents.
    7. Multimodal RAG — Retrieve text, images and tables together.
    8. Agentic RAG — RAG with agents that decide the search strategy.
    9. Tabular Data Retrieval (Text-to-SQL) — Answer questions about databases in natural language.
    10. Spreadsheet and Semi-Structured Data Search — Query tables and CSVs with AI and embeddings.

    Why a skill pack instead of a one-off prompt

    Each skill gives the agent a defined trigger, step-by-step procedure and output format, so results are consistent across sessions and easier to review.

    How to use

    1. Install the pack in your SKILL.md-compatible agent.
    2. Ask for the task in plain language (for example: "use graphs for contextual retrieval").
    3. Review the output against the skill's quality checks before using it.

    Outputs are drafts that need human review; they are not professional, legal or financial advice. Tools, APIs and prices change; check current vendor documentation before relying on version-specific details. Quality depends on the context you give the agent (data, requirements, constraints). Instruction-only skills: they produce documents and plans, not running software.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 8 days ago

    • Free with an Agensi account

    Permissions

    No special permissions declared or detected

    Listed8 days ago

    What's inside

    Frequently Asked Questions

    10 skills, each with SKILL.md and references/sources.md: GraphRAG and Knowledge Graphs; Knowledge Graph Construction; Ontologies and Taxonomies; Knowledge Representation; PDF Ingestion and OCR for RAG; Complex Document Parsing; Multimodal RAG; Agentic RAG; Tabular Data Retrieval (Text-to-SQL); Spreadsheet and Semi-Structured Data Search.

    Projects with PDFs, tables, graphs and semi-structured data.

    Agents that read the SKILL.md format. See the Compatibility note for what was actually tested.

    New in Data & Databases