What is pixie-qa CLAUDE.md?

Project instructions for pixie-qa, a Python package and coding-agent skill for checking AI applications. It covers the project structure, tools, and package management.

Use it when changing pixie-qa code, managing its Python environment with uv, or working with its tests, type checks, formatting, and AI-app monitoring code.

It gives a coding agent the project context needed to work consistently with pixie-qa. This reduces confusion about where code belongs and how to run checks.

pixie-qa CLAUDE.md — Instructions file by yiouli

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods

npx agentmods add instructions/yiouli/pixie-qa/claude-md

Clone the repo

git clone --depth 1 https://github.com/yiouli/pixie-qa

Per session 1,860 This file is loaded in full into every session.

When invoked 1,860 The same file — it is already loaded in full.

Security scan A 0 findings. Scan, not verified.

Origin original No closer match found in the catalogue.

Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

Model	Per session	Once invoked
Fable 5	$0.01860	$0.01860
Opus 5	$0.00930	$0.00930
Sonnet 5	$0.00372	$0.00372
Haiku 4.5	$0.00186	$0.00186

Measured 2d ago against content hash f8cd648c3b64, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

pixie-qa CLAUDE.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

CLAUDE.md · 238 lines

How it starts

The opening of the file, as written. The whole thing — 238 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CLAUDE.md — pixie-qa

Project Overview

pixie-qa is a Python package and coding-agent skill for automated quality assurance of AI applications. Published as pixie, it provides instrumentation, evaluation, and observability primitives for LLM-powered apps.

Technology Stack

Python 3.11+ with type hints
uv for package management, virtual environments, and builds
pytest for testing
mypy for static type checking
ruff for linting and formatting
OpenTelemetry SDK and OpenInference for LLM call instrumentation

Package Structure

pixie/
  __init__.py
  assets/
    index.html           # compiled React scorecard (build artifact, gitignored)
  cli/
    test_command.py      # pixie test entry point
  evals/
    scorecard.py         # scorecard data models + template-based HTML generation
  instrumentation/
    __init__.py          # public API: init(), flush()
    spans.py             # ObserveSpan, LLMSpan, message/content types
    handler.py           # InstrumentationHandler ABC
    wrap.py              # wrap() API for data-object-based tracing
    processor.py         # LLMSpanProcessor (OTel SpanProcessor)
    queue.py             # _DeliveryQueue (background worker thread)
    instrumentors.py     # auto-discovers and activates OpenInference instrumentors
    py.typed

frontend/                # React scorecard SPA source
  src/                   # React components, types, styles
  package.json           # React 19, Vite 6, vite-plugin-singlefile
  vite.config.ts         # builds to ../pixie/assets/

tests/
  pixie/                 # automated tests (pytest)
    cli/
      test_test_command.py   # CLI unit tests
      e2e_fixtures/          # mock evaluators and datasets for e2e
    evals/
    instrumentation/
  manual/                # manual testing fixtures (not run by pytest)
    mock_evaluators.py
    datasets/
      sample-qa.json     # run with: pixie test tests/manual/datasets/sample-qa.json

specs/                   # design specs and architecture docs

Read the full file on GitHub · 238 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

2d ago First seen · 238 lines · 1,860 tokens per session scan A f8cd648c3b64

Subscribe to this mod's changes

pixie-qa CLAUDE.md is an instructions file published in the GitHub repository yiouli/pixie-qa (7 stars, last pushed 4mo ago), licensed MIT. It adds 1,860 tokens to every session, about $0.0093 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Data and AI LLM Languages Testing agent-skills ai-evals dev eval skill

pixie-qa CLAUDE.md

Getting it into your agent

What it costs to keep this loaded

Grade A, and why

Nothing flagged

How it starts

CLAUDE.md — pixie-qa

Project Overview

Technology Stack

Package Structure

What this file has done since we first saw it

Other instructions, from other repositories

avoid-ai-writing CLAUDE.md

rolecraft AGENTS.md

claude-omakase CLAUDE.md

pencil-skill AGENTS.md

skill.color-expert CLAUDE.md

react-flow-skill CLAUDE.md