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 skills/98ping/chip/myopenmathnpx skills add 98ping/chip --skill myopenmathgit clone --depth 1 https://github.com/98ping/chipWhat 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 | $0.00128 | $0.00907 |
| Opus 5 | $0.00064 | $0.00453 |
| Sonnet 5 | $0.00026 | $0.00181 |
| Haiku 4.5 | $0.00013 | $0.00091 |
Grade A, and why
myopenmath 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 — 63 lines — stays where its author put it; the contents beside it link to each section on GitHub.
MyOpenMath homework
Drives a MyOpenMath (IMathAS) assignment end to end with the Claude in Chrome
tools. The full runbook — every trap, with the fix — lives in
tooling/MY_OPEN_MATH.md. Read it before the first interaction; the summary
below is the shape of the job, not a substitute.
The graph reader is tooling/myopenmath-graph.js.
Ask first
Reading the page and computing answers is ordinary help. Typing into the answer boxes and clicking Submit Question each need Max's explicit go-ahead. Ask once, then that covers the run. Never enter credentials anywhere.
Steps
-
Locate — from any Canvas tab, hit the assignments API with a
search_termto get the assignment id and confirm the tool URL ismyopenmath.com(tooling/MY_OPEN_MATH.md§1). -
Launch and identify the mode. This decides everything downstream:
- "Load … in a new browser window" → click it. A new top-level tab joins
your tab group.
read_page,get_page_textandjavascript_toolall work on the real DOM. Much easier — prefer it. - Resume/Start inside the page → cross-origin iframe. Screenshot-only,
the nav arrows wedge, and you need the
about:blankreload trick. See §2.
- "Load … in a new browser window" → click it. A new top-level tab joins
your tab group.
-
Scope the set with
#/print— all questions on one page, real math and graphs. Note which are graph-based. Then reload and Resume. -
Per question (
location.hash = '#/skip/N'):read_pagefor refs, screenshot for layout.- Zoom any radical, exponent or fraction in the question text.
√x + 3vs√(x+3),x²vsx³,s(x)=4/3vss(x)=4— each has bitten. - Graph question → load
myopenmath-graph.jsonce, then__C(n),__at(x),__solve(y),__cls(n,gi). Parse the SVG; do not eyeball the plot. - Fill:
form_inputfortype="text"inputs and<select>s; click + type for MathQuill boxes (baretextboxinread_page).form_inputfails silently on MathQuill. - Verify the rendered value (zoom the box) before submitting — retries are
limited (
↺ Nin the header). - Click Submit Question. Save progress does not count.
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 · 63 lines · 128 tokens per session scan A 3d7be5e9779c
myopenmath is a skill published in the GitHub repository 98ping/chip (2 stars, last pushed 7d ago), licensed MIT. It adds 128 tokens to every session and 907 once invoked, about $0.0006 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 skills, from other repositories
release
发布 Usage4Claude 新版本时使用。当用户说“发布新版本 / 发版 / 出新版 / release / 打 tag 发布 / 准备发版材料”等，用本 skill 引导完成从收集变更、编写 CHANGELOG 与 RELEASENOTES、更新版本号、编译验证，到发版 commit、CI 自动发布的完整流程。.
capture-usage4claude-screenshots
Automate Usage4Claude interface screenshots with CleanShot X on macOS. Use when Codex needs to capture localized Usage4Claude menu-bar popover screenshots, switch Usage4Claude display/language settings, save files such as [email protected], or troubleshoot CleanShot/window-capture automation that depends on…
che-word-mcp
A Swift-native MCP server for Microsoft Word (.docx) document manipulation. Provides 83 tools for reading, writing, and modifying Word documents without requiring Microsoft Word installation.
project-setup
Scaffold a new client project folder from the client-project template, retrofit an existing client folder to the template structure, or reconcile a client folder's memory files after a thread loss. This is the client-engagement variant of the project-setup skill — for internal products, use…
topical-authority-map
Builds a 4-cluster Topical Authority Map from a seed keyword. Outputs an .xlsx spreadsheet with five tabs (Core Pages, AOR Pages, Linking Map, Buyer Journey, Publishing Order) plus a markdown strategy document. Interactive, asks the user to confirm central entity, source context, persona, classification, and gaps at…
dream
Consolidate and reorganize project memory (MEMORY.md) by invoking the dream-consolidator named subagent (Sonnet, pinned via the agent's model: frontmatter — NEVER pass a model: parameter yourself, it defeats the pin and runs the dream on expensive Opus). The subagent prompt is a CLOSED TEMPLATE: send its fields and…