xerj-code

Featured

Reference-coding with XERJ. Clone the libraries that already solved your problem, index them locally, and retrieve the exact implementation before writing code — so the agent reads passages instead of re-deriving algorithms across retry loops. Use when starting a task in an unfamiliar API, porting an algorithm, or when you have already looped twice on the same error.

AI & Automation 19,231 stars 1445 forks Updated today Apache-2.0

Install

View on GitHub

Quality Score: 93/100

Stars 20%
100
Recency 20%
100
Frontmatter 20%
70
Documentation 15%
100
Issue Health 10%
50
License 10%
100
Description 5%
100

Skill Content

# XERJ.code — reference coding An agent that does not know an API guesses, runs, fails, and guesses again. Each lap costs **output** tokens, which are the expensive kind. This skill replaces laps with a lookup: clone the repositories that already contain a correct implementation, index them with XERJ, and retrieve the passage before writing. The trade is deliberate. Retrieval spends **input** tokens, which are cheaper per token and cacheable; loops spend **output** tokens. See [`COSTS.md`](../../docs/case-studies/reference-coding/COSTS.md) for the arithmetic and for the conditions under which this trade **loses**. > **Status, stated plainly: measured, and the value is real but narrow.** > (Full case study with real tokens and dollars: [`docs/case-studies/reference-coding/CASE_STUDY.md`](../../docs/case-studies/reference-coding/CASE_STUDY.md), 2026-08-05.) > > The comparison that matters is against **native Claude Code** — the same agent > with tools that would grep the source itself — not a bare model. Measured across > three regimes, same agent, one objective verdict (compiles + passes hidden tests): > > - **Unfamiliar code with a contract the model can't recall** (a seal, a > generational handle, a lazy refill, a specific hash scheme — measured across > **seven purpose-built libraries, 21 runs**): native and xerj both solve > **21/21**; **xerj uses 1.5× fewer output tokens and costs 1.3× less than > native**, because retrieval replaces the grep. Pure memory **fai...

Details

Author
xerj-org
Repository
xerj-org/xerj
Created
4 months ago
Last Updated
today
Language
Rust
License
Apache-2.0

Integrates with

Similar Skills

Semantically similar based on skill content — not just same category

AI & Automation Featured

ripwire-efficient

A token + accuracy DISCIPLINE for ANY read, not a moment: map before you open files. Reach for it the instant you catch yourself about to open more than ~2 files to figure something out — anywhere in a task. Also the moment you catch yourself typing "let me search the codebase", "let me read that file", or "let me look at a few files first": that sentence IS the trigger. File reads dominate an agent's token cost, and less context is measurably MORE accurate, not just cheaper; a small ranked map beats a fan-out of whole-file reads on both. Code-repair accuracy fell 29% → 3% as context grew 32K → 256K tokens (LongCodeBench), so "read a few more files to be safe" is the instinct this replaces, not a safe default. The reflex is one line: run the cheapest verb that answers the question, then read only the 2-3 files it ranks highest. Fires ALONGSIDE the moment skills (orient, navigate, change-check…), never instead of them. Backed by ripwire (deterministic, on PATH).

2,552 Updated today
redhat-et
AI & Automation Solid

xerj

Find current XERJ documentation and API contracts for folder indexing, code retrieval, search, and agent memory. Use when integrating or operating a user-selected XERJ node.

19,231 Updated today
xerj-org
AI & Automation Listed

agent-efficiency-retrofit

Analyze recent Claude Code and Codex CLI conversations for repeated inefficient tool calls, token waste, retry loops, avoidable errors, and recurring setup friction; rank improvements by benefit/ease, then implement the smallest durable guardrail after user approval.

5 Updated today
kookr-ai