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.
npx agentmods add instructions/crypdick/pynchy/agents-mdgit clone --depth 1 https://github.com/crypdick/pynchyWhat 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 | $0.01705 | $0.01705 |
| Opus 5 | $0.00852 | $0.00852 |
| Sonnet 5 | $0.00341 | $0.00341 |
| Haiku 4.5 | $0.00170 | $0.00170 |
Grade A, and why
pynchy AGENTS.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 yesterday.
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.
How it starts
The opening of the file, as written. The whole thing — 83 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Pynchy
Personal Claude assistant. See README.md for philosophy. See installation guide for installation. See architecture for architecture. See CONVENTIONS.md for design principles (composition over inheritance, parse-don't-validate, semantic types, code/doc coupling) — apply them when writing or reviewing code.
Architectural Direction
Pynchy's architectural ambition is a modular monolith with semantic domain contracts, use-case-owned ports, concrete adapters, and explicit composition roots—not a distributed-services or dependency-injection-framework rewrite. Before changing a cross-subsystem dependency, read the service boundary guide. Treat architecture.toml as the executable source of truth for current package ownership, public surfaces, and allowed role dependencies.
Burn down architecture-baseline.toml package by package. For each multi-module
package with cross-package consumers, migrate its sole public surface to
<package>.api; a façade made entirely of curated re-exports is valid. Move all
consumers in the package pass, update public_modules, and remove every stale
baseline entry. Correct package roles when direction is wrong. Do not weaken the
policy or regenerate the baseline merely to reduce the count.
Deployment Awareness
You are usually NOT running on the production host. The live host is deployment-specific and should come from local memory, environment, or the operator, not from public repo defaults. Before making changes that affect the live service (config.toml, server-side files, service restarts), read the pynchy-ops skill for deployment procedures, auto-deploy behavior, and how to observe the running service.
Quick Context
Python process that connects to messaging channels (WhatsApp, Slack, etc. via plugins), routes messages to Claude Agent SDK running in containers (Apple Container on macOS, Docker on Linux). Each group has isolated filesystem and memory.
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.
- yesterday First seen · 83 lines · 1,705 tokens per session scan A af478b5bf8df
pynchy AGENTS.md is an instructions file published in the GitHub repository crypdick/pynchy (10 stars, last pushed 7d ago), licensed MIT. It adds 1,705 tokens to every session, about $0.0085 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.
Other instructions, from other repositories
nanoclaw CLAUDE.md
Claude Code instructions for nanocoai/nanoclaw, covering ⚠️ stop — read this first if you are claude ⚠️, nanoclaw, quick context, entity model and two-db session split.
nanoclaw AGENTS.md
AGENTS.md instructions for nanocoai/nanoclaw, a project described as: A lightweight alternative to OpenClaw that runs in containers for security. Connects to WhatsApp, Telegram, Slack, Discord, Gmail and other messaging apps,, has memory, scheduled jobs, and runs directly on Anthropic's Agents SDK.
ChatClaw AGENTS.md
AGENTS.md instructions for zhimaAi/ChatClaw, covering chatclaw agents.md, 项目概述, 通用开发规范, 语言约定 and codex superpowers.
openclaw-config AGENTS.md
AGENTS.md instructions for TechNickAI/openclaw-config, covering project context for ai assistants, project overview, tech stack, project structure and code conventions.
openclaw-config CLAUDE.md
Claude Code instructions for TechNickAI/openclaw-config, a project described as: Give your AI assistant memory, skills, and autonomy. Persistent memory, integration skills, and autonomous workflows.
clawe CLAUDE.md
Claude Code instructions for getclawe/clawe, covering clawe, commands, structure, data layer and environment variables.