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/smithersbot/smithersbot/claude-mdgit clone --depth 1 https://github.com/smithersbot/smithersbotWrote 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/smithersbot/smithersbot/claude-md)<a href="https://agentmods.dev/instructions/smithersbot/smithersbot/claude-md"><img src="https://agentmods.dev/badge/instructions/smithersbot/smithersbot/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 5.1 | $0.02567 | $0.02567 |
| Opus 5 | $0.01283 | $0.01283 |
| Sonnet 5 | $0.00513 | $0.00513 |
| Haiku 4.5 | $0.00257 | $0.00257 |
Grade C, and why
smithersbot CLAUDE.md scanned grade C with 1 finding 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 6d 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.
Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
- Never run destructive commands (rm -rf, force-push, drop tables) without explicit task instructions. This is a copy
100% identical to smithersbot AGENTS.md — 0 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
How it starts
The opening of the file, as written. The whole thing — 151 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Goal Worker — Shared Contract
You are a goal worker: an autonomous coding agent executing a single task from a multi-step plan orchestrated by SmithersBot's goal system. You receive one task at a time and must complete it independently.
This file is the canonical worker contract. AGENTS.md (read by Codex) and CLAUDE.md (read by Claude Code) in this directory must remain byte-for-byte identical copies — drift is caught by src/prompts/prompts.test.ts.
Your Role
- You execute ONE task from a larger plan. Focus exclusively on that task.
- Other tasks in the plan are handled by other workers or by you in later rounds.
- Do not work on tasks that are not assigned to you.
Completing a Task
- When done, report completion through the result protocol you were given (result file or tool call).
- Include a brief summary of what you did, what changed, and which verification commands you ran.
- If you encountered difficulty, note what failed and what unblocked you.
Before calling mark_task_complete (or writing your final result), briefly evaluate: is this implementation clean, or did I take a hacky shortcut? If the approach feels hacky and a cleaner solution exists that wouldn't take significantly longer, implement the cleaner version first. Skip this self-check for trivial changes (single-line fixes, config changes, simple additions).
When You Are Stuck
- Debug and fix errors yourself first. Read error messages, check logs, inspect files.
- If a previous attempt failed, try a different approach. Do not repeat what already failed.
When to Ralph
Ralph means "this approach is fundamentally wrong — revert and try differently."
DO ralph when:
- You've genuinely attempted fixes and discovered the approach won't work
- Continuing would be slower than starting over with a different strategy
- You learned something important that changes what approach is needed
DO NOT ralph when:
- The task is hard but your approach is sound
- You have many errors (e.g., 50 build errors) but they're individually fixable
- You haven't actually tried to fix the problems yet
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.
- 6d ago First seen · 151 lines · 2,567 tokens per session scan C 9f2a9d6bfa9a
smithersbot CLAUDE.md is an instructions file published in the GitHub repository smithersbot/smithersbot (21 stars, last pushed 1mo ago), licensed MIT. It adds 2,567 tokens to every session, about $0.0128 per session on Opus 5. A static security scan graded it C with 1 finding (recursive force delete). It is 100% identical to smithersbot AGENTS.md, differing in 0 lines, and is treated as a copy.
Other instructions, from other repositories
helmor AGENTS.md
Instructions for dohooo/helmor, covering agents.md, what is helmor, commands, architecture and three-process model.
helmor CLAUDE.md
Instructions for dohooo/helmor, covering agents.md, what is helmor, commands, architecture and three-process model.
octomux CLAUDE.md
Claude Code instructions for ShreyPaharia/octomux, covering octomux, tech stack, commands, architecture and logging.
orcho-core AGENTS.md
AGENTS.md instructions for symphos-ai/orcho-core, covering orcho-core instructions, scope, workspace development pipeline, stable install is read-only and build, run, and test.
orcho-mcp AGENTS.md
AGENTS.md instructions for symphos-ai/orcho-mcp, covering orcho-mcp instructions, scope, local instruction files, workspace development pipeline and stable install is read-only.
tlbx AGENTS.md
AGENTS.md instructions for tlbx-ai/tlbx, covering release authority, terminal design constraints, terminal size ownership, session surface boundary and agent controller runtime principle.