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/rcrsr/claude-code-runner/claude-mdgit clone --depth 1 https://github.com/rcrsr/claude-code-runnerWhat 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 | $0.00716 | $0.00716 |
| Opus 5 | $0.00358 | $0.00358 |
| Sonnet 5 | $0.00143 | $0.00143 |
| Haiku 4.5 | $0.00072 | $0.00072 |
Grade A, and why
claude-code-runner 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.
How it starts
The opening of the file, as written. The whole thing — 59 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
@README.md
Commands
npm run check # Full validation: typecheck + lint + format + test
npm test # Run all tests
npm run build # Compile TypeScript
Architecture
CLI modes: prompt (single execution), command (templates from .claude/commands/), skill (slash command /<name>), script (Rill workflows), docs (print rill/CCR reference)
Rill runner (src/rill/runner.ts): Executes .rill scripts using the Rill interpreter with CCR host functions.
Host functions (provided to Rill via ccr:: namespace):
ccr::prompt(text, model?, timeout?)→ Execute Claude promptccr::command(name, args?, timeout?)→ Run command template (args is list)ccr::skill(name, args?, timeout?)→ Run slash command (args is list)ccr::state(text)→ Set the script status line textccr::get_result(text)→ Extract the last XML result from outputccr::has_result(text)→ Check if text contains ccr:resultccr::file_exists(path)→ Check file existenceccr::get_frontmatter(path)→ Parse YAML frontmatterccr::has_frontmatter(path)→ Check if file has frontmatter
Script variables: $ARGS (positional args list, $ARGS[0]…), $ENV (env vars), $name (named args frontmatter), $varname (captures via => $varname). Note: $1/$2/$ARGUMENTS are .claude/commands/ and skill template variables, not valid in .rill scripts.
Constants: All magic numbers live in src/utils/constants.ts
Code Conventions
- Barrel exports: Each module directory has
index.tsre-exporting public APIs - Factory functions over classes: Use
create*functions (e.g.,createLogger()) as constover enums: Use const objects with derived types- Type guards with unions: Place
is*type guards next to discriminated union definitions - No global state: Pass state objects through function calls (
FormatterState,RunnerContext)
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 · 59 lines · 716 tokens per session scan A c7eba6b1eeab
claude-code-runner CLAUDE.md is an instructions file published in the GitHub repository rcrsr/claude-code-runner (5 stars, last pushed 2mo ago), licensed MIT. It adds 716 tokens to every session, about $0.0036 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
mcp-server-powerpoint mcp-server-guide.instructions.md
Instructions for sbroenne/mcp-server-powerpoint, covering mcp server development guide, llm-facing content rules, two kinds of tools: hand-written vs. generated, implementation pattern: single hand-written dispatch tool and error handling (mandatory).
mcp-server-powerpoint code-review.instructions.md
Instructions for sbroenne/mcp-server-powerpoint, covering copilot code review, interop and resource safety, contract and path completeness and protocol, tests, and user-facing surfaces.
ai-workflows AGENTS.md
Instructions for dryvist/ai-workflows, covering ai-workflows, architecture, directory structure, workflow types and consumer repo caller pattern.
1ai-skills CLAUDE.md
Instructions for oyi77/1ai-skills: This is a Claude Code project configuration for a 1ai-ecosystem repository.
babysitter CLAUDE.md
Claude Code instructions for a5c-ai/babysitter, covering claude.md, repo-specific overrides, read these first, quick commands and claude code local notes.
initrunner CLAUDE.md
Instructions for vladkesler/initrunner, covering initrunner and project map.