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/tradebaas/groundwork/ingestnpx skills add Tradebaas/Groundwork --skill ingestgit clone --depth 1 https://github.com/Tradebaas/GroundworkWhat 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 | $0.00094 | $0.00869 |
| Opus 5 | $0.00047 | $0.00434 |
| Sonnet 5 | $0.00019 | $0.00174 |
| Haiku 4.5 | $0.00009 | $0.00087 |
Grade A, and why
ingest 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 3d 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 — 62 lines — stays where its author put it; the contents beside it link to each section on GitHub.
ingest: turn documents into Markdown before they cost tokens
Reading a PDF, Office file, or image directly is expensive and lossy: the bytes are not text, so an agent burns tokens on encoding noise or cannot read them at all. markitdown (https://github.com/microsoft/markitdown, MIT, Microsoft) converts them to clean Markdown that carries the structure (headings, tables, lists) at a fraction of the tokens. Convert first, then read the Markdown.
Run the decision ladder first. Convert only the file the task needs, only the pages or sheets that matter, and read the result the same way you read any file: the part that answers the current question, not the whole thing. Do not bulk-convert a folder "to have it".
When to use
- A task needs what is inside a PDF, DOCX, PPTX, XLSX, image, audio file, HTML page, or data file.
- The owner hands over brand or reference material for
docs/design/reference/or a spec. - The product being built must accept and parse user-uploaded documents (then markitdown, or its Python API, becomes part of the product, not just a build-time helper).
Install and invoke (from the living source)
Python with pip is the prerequisite (minimum version per the README); Node stays Groundwork's
only always-on requirement, so this is opt-in, installed only when you actually convert
something. Take the exact syntax from the markitdown README (linked above) at the moment of
use: the install command with optional per-format extras, the CLI, the markitdown-mcp server
(one convert_to_markdown tool, for tools that speak MCP) and the Python MarkItDown API all
live there. This skill deliberately embeds none of those commands: frozen copies drift from
upstream (skill-author rule: no code snippets that go stale, point at the living source).
It converts: PDF, PowerPoint, Word, Excel, images (EXIF + OCR), audio (EXIF + transcription), HTML, CSV, JSON, XML, ZIP (walks the archive), YouTube URLs, EPub.
Where the output goes
- Throwaway conversions (you just need to read something once): write to a scratch/temp path, never commit it.
- Durable reference the project should keep:
docs/design/reference/or the relevant spec folder, and add the manifest row if it lives underdocs/(docs/README.md rules). - Do not commit the Markdown of a copyrighted or personal-data source without clearing it through
comply. Converting does not change who owns the content.
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.
- 3d ago First seen · 62 lines · 94 tokens per session scan A 07f3c82aca87
ingest is a skill published in the GitHub repository Tradebaas/Groundwork (2 stars, last pushed 7d ago), licensed MIT. It adds 94 tokens to every session and 869 once invoked, about $0.0005 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
pdf-doc-processing
Use when reading, extracting, summarizing, or converting PDFs related to landscape project documentation, client briefs, reference packs, estimates, contracts, or scanned drawings.
skill-doc-delivery
Convert markdown to DOCX, PPTX, XLSX, PDF office documents — use when you need exportable deliverables.
document-gen-resilient
Multi-path document generation with tool checks, Unicode handling, and Python fallbacks.
pdf-explore
Read, extract, and cross-check content across scientific PDFs. Use when a task needs methods, figures, tables, citations, accessions, or claims from multiple places in one or more papers.
make-resume
中文可编辑简历制作技能：根据用户经历选择或复刻模板，生成可编辑 HTML 简历并提供 PDF 导出；当用户输入“/make-resume”或要求制作、修改、复刻简历文件时使用。.
paper-reader
Use when user asks to "read paper", "analyze paper", "summarize paper", "读论文", "分析文献", "帮我看一下这篇paper", "论文笔记", or provides a PDF file that appears to be an academic paper. Specialized for CV/DL papers. Also supports Zotero integration: "读一下这篇论文 ...", "快速看一下这篇论文 ...", "批判性分析这篇论文 ...", "读一下 Zotero 里的 XXX", "批量读一下 Zotero…