Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

Get Started Free →

Instructions — measured AI agent skill | DecimalAI

docxology/LLM Extraction & KG Construction

SkillScore91Maintained

Core inference logic, prompt construction, and RDF/Nanopublication serialization.

by docxology

Links this skill to your agents and keeps it current with the author’s updates.

Automatic

AI Engineer pack

decimalai skills pull docxology-llm-extraction-kg-construction

Get this skill — curl · raw file

curl -fsSL https://app.decimal.ai/s/docxology-llm-extraction-kg-construction/SKILL.md

Raw file: /s/docxology-llm-extraction-kg-construction/SKILL.md ↗ · machine-readable summary at /s/docxology-llm-extraction-kg-construction.json ↗

In productionGlobal

how this skill behaves on real traffic · last 30 days

Runs

—

no usage data yet

Used when offered

—

offered 330 · activation not measurable for prompt-injection usage

Anonymized aggregate across every org using this skill — proof from real production use, not a demo.

Does this skill help? · Benchmark: with vs. withoutsign in to run this on your agents

22 test cases from the publisher’s eval.yaml · each run twice on gemini-3.6-flash — with the skill injected, then without · last run 6d ago

Yes — +45% lift-24% tokens0% turns91% of whole cases pass with skill (20/22)

Test case	Without → With	Effect	Δ tokens	Δ turns
case-16	✗→✓	▲ Improved	-29%	0%
case-17	✗→✓	▲ Improved	-25%	0%
case-02	✗→✓	▲ Improved	-4%	0%
case-03	✗→✓	▲ Improved	-28%	0%
case-06	✗→✓	▲ Improved	-24%	0%

Showing 5 of 22 cases

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v4 — 8/1/2026

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

docxology / template

projects/templates/template_literature_meta_analysis/src/knowledge_graph/SKILL.md

View on GitHub ↗

License Apache-2.0Imported 8/1/2026Version v1

Skill contents

SKILL.md

Instructions

You are interacting with the src/knowledge_graph/ module. This handles the transition from natural language text to structured assertions via Large Language Models.

Agentic Interface (MCP Strategy)

Config-Driven Orchestration: Never hardcode LLM parameters (models, temperature, minimum confidence thresholds) within Python scripts. Bind them tightly to LLMConfig structures derived securely from config.yaml.
Robust Categorization: Respect strict boundaries on confidence gating. When crafting or modifying prompt templates, ensure the extraction schema firmly permits "irrelevant", "neutral", and "don't know" states to neutralize base-model hallucination tendencies.
Streaming Persistence: Assertions must emit safely to disk via incremental JSON Lines or valid TriG records. High-volume document processing should not accumulate global lists in RAM.

Want to see if docxology/LLM Extraction & KG Construction would activate on your phrasing? Try it before you install — no LLM call, no signup.

Other measured skills in the registry, with their headline benchmark lift.