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/mrclrchtr/supi/claude-mdgit clone --depth 1 https://github.com/mrclrchtr/supiWrote 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/mrclrchtr/supi/claude-md)<a href="https://agentmods.dev/instructions/mrclrchtr/supi/claude-md"><img src="https://agentmods.dev/badge/instructions/mrclrchtr/supi/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 | $0.04617 | $0.04617 |
| Opus 5 | $0.02308 | $0.02308 |
| Sonnet 5 | $0.00923 | $0.00923 |
| Haiku 4.5 | $0.00462 | $0.00462 |
Grade A, and why
supi 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 today.
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 — 194 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
SuPi (Super Pi) is a curated extension repo for the pi coding agent (@earendil-works/pi-coding-agent).
It is a pnpm workspace monorepo of installable pi extensions. pi loads the extensions directly as TypeScript — there is no build step.
Development status
SuPi is pre-release and not API-stable. Intentional breaking changes to package APIs, commands, configuration formats, and extension behavior are allowed when they improve the design. Do not treat backwards compatibility as a blocker unless a task explicitly asks for it.
Pi docs-first rule
- Never assume pi harness APIs, behavior, or conventions from memory or model priors.
- Before changing code or giving guidance about pi-specific behavior, read the relevant installed pi docs (
README.md, matching files indocs/, andexamples/) and follow linked.mdcross-references. - Start with
docs/index.mdfor getting an overview of PI's docs. - For Pi context mechanics verified in this repo (request payload, prompt caching, compaction, extension hooks for context control), see
docs/pi/context-architecture.md.
Documentation expectations
docs/INDEX.mdis the docs index — start there for repo documentation; add new docs per its maintenance rules.- Add JSDoc for exported APIs, config surfaces, and non-obvious behavior; skip boilerplate for trivial private code.
- Add inline JSDoc for complex internal logic where a short explanation would help maintainers.
Package layout convention
- Follow
docs/conventions/package-layout.mdfor repo-wide package structure. - Standardize package boundaries with
src/api.ts,src/index.ts, andsrc/extension.tswhen the package role requires them. - Prefer package-level tests in
__tests__/unit/and__tests__/integration/, with__tests__/helpers/and__tests__/fixtures/as needed. - Prefer domain folders over catch-all names like
core/,shared/, ormisc/. - Keep small packages flat; add
config/,ui/,session/, or other domain folders only when the package size and responsibilities clearly justify them. Packages that register tools use the per-tool directory layout undersrc/tool/(docs/conventions/tool-architecture.md§ Per-tool directory layout). - Anchor examples: the large-package model (hybrid with domain folders) and the standard package model (flat with package-level tests).
- This convention is the default for new packages and for existing packages when they receive structural work.
- Packages that should stay flat unless they grow:
supi-bash-timeout,supi-context,supi-debug,supi-test-utils. supi-skill-patchesis a private maintenance package. It owns upstream synchronization and per-file compatibility patches; generated skills live in the rootskills/catalog, and selected project-local skills link into.pi/skills/.supi-webshould stay mostly flat, but may usesrc/tool/for per-tool guidance files.
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.
- today Changed · +28 tokens per session 4664e451b8da
- 5d ago First seen · 194 lines · 4,589 tokens per session scan A c2402fcc2c31
supi CLAUDE.md is an instructions file published in the GitHub repository mrclrchtr/supi (86 stars, last pushed today), licensed MIT. It adds 4,617 tokens to every session, about $0.0231 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
delegate-skills AGENTS.md
AGENTS.md instructions for amElnagdy/delegate-skills, covering working on delegate-skills, vocabulary, conventions, before publishing a change and local claude code config.
Albatross AGENTS.md
Instructions for morganlinton/Albatross, covering agent instructions, general and releases.
pi-agent-browser-native AGENTS.md
AGENTS.md instructions for fitchmultz/pi-agent-browser-native, covering pi-agent-browser-native, product-specific guidance, documentation placement, upstream capability baseline and command reference and tool result categories.
baro CLAUDE.md
Claude Code instructions for jigjoy-ai/baro, covering claude.md, project overview, tech stack, directory structure and build commands.
baro AGENTS.md
AGENTS.md instructions for jigjoy-ai/baro, covering claude.md, project overview, tech stack, directory structure and build commands.
pi-codex-goal AGENTS.md
Instructions for fitchmultz/pi-codex-goal, covering pi-codex-goal — agent notes, local pi install policy, verify before finishing and layout.