skills.rest
arpitarya
arpitaryaCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

fux-fetch-rules

Extracts actionable rules from URLs, PDFs, and text files into durable Fux rule entries.

1|Updated Jun 1, 2026
Search Tags:#automation#documentation#compliance#pdf parsing#rule extraction#knowledge index#fux
One-click install
npx skills add https://github.com/arpitarya/fux --skill fux-fetch-rules-arpitarya
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: fux-fetch-rules
Source: https://github.com/arpitarya/fux/tree/main/archive/v0.1/fux/data/skills/fetch-rules
Command: npx skills add https://github.com/arpitarya/fux --skill fux-fetch-rules-arpitarya

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires fux-engine, pypdf.

What problem does it solve? Manually reading long documents (regulations, specs, standards) and transcribing every constraint into project rules is slow and error-prone. This Skill automates the derivation: it reads a source document and converts every actionable constraint, formula, or decision into structured, durable .fux/rules/ entries. ## Core Features & Use Cases - Automatic rule derivation: Parses a URL, PDF, or text file and classifies each finding into typed entries (formula, rule, convention, regulatory, adr, runbook, edge-case, glossary). - Provenance tracking: Every entry records its source: origin and a mandatory **Why:** rationale, validated by fux lint. - Index integration: Builds the Fux index after creation so rules are retrievable later via fux recall. - Use Case: Point it at a SEBI regulation PDF or an internal API spec, and it produces entries like stcg-equity-rate (formula) or date-format-iso8601 (convention), each with scope, edge cases, and citations. ## Quick Start Ask the agent to run /fux fetch-rules on a regulation PDF or spec URL and derive all actionable rules into the project's .fux/rules directory.

Frequently Asked Questions about fux-fetch-rules

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I extract rules from a PDF or URL automatically?▼

Run /fux fetch-rules with the source path or URL. The agent reads the entire document, classifies each actionable constraint into typed entries (formula, rule, regulatory, etc.), and writes them as durable .fux/rules/ files with source provenance.

What types of rules can be derived from a document?▼

Eight types are supported: formula, rule, convention, regulatory, adr, runbook, edge-case, and glossary. The classification is driven by textual signals such as 'must/shall' phrasing, numeric formulas, or step-by-step procedures.

Does fetch-rules support remote PDF files?▼

Yes, remote PDF URLs are handled by the CLI extractor if the pypdf extra is installed via pip install 'fux-engine[pdf]'. Otherwise, download the PDF with curl first and read the local file directly.

Why does fux lint report no-why warnings after deriving rules?▼

Every created entry must include a non-empty **Why:** section explaining the obligation or rationale. Entries missing it fail lint because a rule without its reasoning is considered incomplete; fix all warnings before finishing.

What are the limitations of automatic rule extraction?▼

Pure narrative or background text without obligations is skipped, and vague preferences like 'try to' are excluded unless clearly required. The derivation quality depends on the agent's reading, so entries should be reviewed for faithfulness to the source.

Related Skills

Explore More Software Engineering Skills→
L3DigitalNetL3DigitalNet
6

wiki-sync

Synchronizes GitHub wiki content by cloning, diffing, generating, and pushing pages.

Official
Advanced
matochumatochu
4

ctx

Create feature-specific context documents with bidirectional links to feature files.

Community
Intermediate
itallstartedwithaideaitallstartedwithaidea
31

knowledge-base-rag

Ingest private documents into a vector index and generate cited answers.

Community
Advanced
addxaiaddxai
42

doc-writing

Separate human-authored judgments from AI-generated content using HWPR and AWOR markers.

Official
Intermediate