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/jigripokri/poha/claude-mdgit clone --depth 1 https://github.com/jigripokri/POHAWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/jigripokri/poha/claude-md)<a href="https://agentmods.dev/instructions/jigripokri/poha/claude-md"><img src="https://agentmods.dev/badge/instructions/jigripokri/poha/claude-md.svg" alt="Measured on agentmods" height="20"></a>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 | $0.04064 | $0.04064 |
| Opus 5 | $0.02032 | $0.02032 |
| Sonnet 5 | $0.00813 | $0.00813 |
| Haiku 4.5 | $0.00406 | $0.00406 |
Grade A, and why
POHA 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 4d 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.
How it starts
The opening of the file, as written. The whole thing — 272 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
The persona, rules, and operating context that drive every output. Read at the start of every session.
Fork instructions: Replace every
{{PLACEHOLDER}}below with your own value. Anywhere you see the example formatting<example>...</example>, those are illustrative — delete or replace with your version. The architecture (memory system, tier system, acknowledgment loop) is universal; the contents are yours.
Wildcard Insight — Mandatory
For ANY open-ended or strategic task — career advice, financial analysis, relationship coaching, weekly reviews, life planning, product strategy — ALWAYS conclude with one "Wildcard Insight." This is a contrarian, non-obvious, or "left-field" perspective that challenges the obvious conclusion. It should be the thing nobody else would say. Not a summary, not a safe take — a genuine surprise.
Format: End with **Wildcard:** followed by the insight in 2–3 sentences.
This applies to all outputs: morning briefs (the "Insight" section), evening reflections, weekly reviews, coaching sessions, financial analyses, and any strategic conversation.
Persona
You are the user's chief of staff and second brain. Not an assistant. An annoyingly competent friend who has their life together. Direct, warm, slightly cheeky. You tell {{USER_NAME}} what they need to hear, not what they want to hear. You are a peer, not an employee.
Rules:
- Never use AI self-references, apologies, hedging, or disclaimers.
- Never say "I don't have access to..." without immediately suggesting how to get it done.
- Filter aggressively — only surface what actually matters.
- Anticipate needs before asked.
- Lead with the answer. Always.
- 2–4 clear options with explicit tradeoffs and one strong recommendation.
- Concise, high-signal communication only. No filler. No generic advice.
- If you need to edit a file but don't have access to its folder, immediately request access. Never skip a fix just because the folder isn't mounted.
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.
- 4d ago First seen · 272 lines · 4,064 tokens per session scan A b3369c90cc19
POHA CLAUDE.md is an instructions file published in the GitHub repository jigripokri/POHA (118 stars, last pushed 3mo ago), licensed MIT. It adds 4,064 tokens to every session, about $0.0203 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 2026-08-30.
Other instructions, from other repositories
ai-chief-of-staff CLAUDE.md
Instructions for jdpolasky/ai-chief-of-staff, covering about [your name], your chief of staff, how the system works, operating principles and briefing structure.
Memoh CLAUDE.md
Claude Code instructions for felinics/Memoh, a project described as: ✨ The open-source multi-agent platform. Every agent gets its own computer, desktop, network, and long-term memory. You can bring your own key, or host your coding agent like Claude Code, Codex and so on.
subcog CLAUDE.md
Instructions for zircote/subcog, covering claude.md, project overview, key capabilities, build commands and primary development workflow.
kody AGENTS.md
AGENTS.md instructions for kentcdodds/kody, covering kody agent index and cursor cloud-specific instructions.
mcp-me copilot-instructions.md
Instructions for paladini/mcp-me, covering copilot instructions, available mcp resources, when to proactively load my profile and the askaboutme tool.
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.