fux

Manages project knowledge as frontmatter rules with deterministic index, graph, and memory views.

1|Updated Jun 1, 2026
One-click install
npx skills add https://github.com/arpitarya/fux --skill fux-arpitarya
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: fux
Source: https://github.com/arpitarya/fux/tree/main/archive/v0.1/fux/data/skills/agents
Command: npx skills add https://github.com/arpitarya/fux --skill fux-arpitarya

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Agent-assisted projects repeatedly re-derive business rules, conventions, and cross-session memory from scratch each session. This Skill stores each fact once as a frontmatter entry in a .fux/ directory and serves it back through a generated index, graph, and memory views, with deterministic maintenance commands that cost nothing in LLM calls. ## Core Features & Use Cases - Unified knowledge substrate: Rules, glossary, ADRs, memory, and narrative docs live in one frontmatter format, with INDEX.md, rules.json, and graph views regenerated by fux build. - Drift detection and validation: fux check flags dead code references, stale rules, and conflicts; fux verify runs invariant assertions so rules that drift from code fail rather than warn. - Deterministic retrieval and governance: fux recall retrieves relevant rules by keyword, fux ratify records tamper-evident rule approvals, and fux gate enforces knowledge health in CI or pre-commit hooks. - Use Case: A team documenting why a pricing formula works a certain way creates a formula rule with code_refs pointing at the implementation lines; later sessions answer questions via fux recall instead of re-reading the codebase. ## Quick Start Ask the agent to run /fux init in your project to scaffold the .fux knowledge directory, then use /fux recall followed by your question to retrieve the relevant rules.

Frequently Asked Questions about fux

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

FAQPage Schema
How do I set up persistent project knowledge for AI agents?▼

Run /fux init in your project to scaffold a .fux directory with rules, glossary, and memory folders plus a config file. Then create entries with /fux new and regenerate the index with /fux build so agents read the index instead of re-scanning code.

How do I detect when documentation drifts from code?▼

Run fux check to flag dead code references, stale rules detected via git history, and conflicting entries. The fux verify command additionally executes invariant assertions so a rule that no longer matches the code fails rather than just warning.

Does fux require an LLM or API key to run?▼

No. All core commands such as build, check, recall, verify, and gate are deterministic shell, AST, and parse operations with zero API cost. Only the plan, adr, and debate skills use the LLM session you are already running.

What Python version does fux require?▼

Fux requires Python 3.11 or newer because it relies on tomllib for configuration parsing. The skill's setup step resolves an available 3.11+ interpreter and installs the engine if the fux binary is not already present.

Can fux enforce knowledge rules in CI pipelines?▼

Yes. The fux gate command runs in CI and exits with code 2 on blocking findings, and fux gate --install wires it as a git pre-commit hook. Strictness levels from off to strict are configured per project in .fux/config.toml.