Independent by design · no pay-to-rank placements  · How we evaluate tools →
Anclum Submit
M

mcnamee/claude-skills

My Claude Skills for Productivity

Before you install: no copy-paste command · no license recorded · link is dead

First step: | Plugin | Version | What it does | pip install |

View on GitHub (unreachable when checked on 2026-11-12)This link returned 404 on 2026-11-12
Star count
1 ★ · mcnamee/claude-skills · read from the GitHub search API · (checked 2026-10-26 20:00:26)
Source
read directly from its GitHub page · type: open-source project
Harvest history
first seen 2026-10-26 · last seen 2026-11-14 · seen 23× · listed by 1 sources
Score
score 103 · 1 ★ · 1 sources · repo
FAQ (answers come from our records)
Who is it for?
We filed it under these tags: 开源项目. The tags come from the entry itself, not a taxonomy we imposed.
Who maintains it?
repository mcnamee/claude-skills;first catalogued 2026-10-26;last seen 2026-11-14。There is no official certification or quality rating here — we record sources and dates only.

# Claude Skills

以下摘录自该仓库的 README(片段) · 在 GitHub 读完整 README →

Three things, all for Claude Code, all working entirely offline:
plugins/ — eight MCP servers that give an agent hands on the
things enterprise work actually lives in: Word documents, Excel workbooks,
Outlook mail, Confluence, Jira, PDFs, plus a local RAG knowledge base to tie
them together. This repo doubles as a plugin marketplace for them.
eva/.claude/skills/ — standalone skills that
need no server and no install beyond the eva/ copy below.
eva/.claude/agents/ — subagents that take a whole
job away and hand back a finished result, built on top of the servers and
skills above.
Plus eva/ — the working folder they all read, write and index,
carried here as a scaffold you copy to H:\Eva. The skills and agents live in
its .claude\ folder, so they install with it and stay scoped to H:\Eva
rather than your whole account. It holds the knowledge base,
the document library, and the templates — the blank .docx/.pptx files
new documents and decks are built from. It also carries
eva/CLAUDE.md, the standing instructions that make the
assistant Eva — an executive virtual assistant who writes in Australian
English, without em dashes or the other AI tells, and never fills a gap in the
source material with something invented.
> Using OpenCode instead of Claude Code? The same servers and skills work
> there too - see OPENCODE.md for the install.
## Why this exists
Built for Windows endpoints in an Enterprise environment. No internet calls, no telemetry,
no registry lookups. Copy the folder across, add it as a local marketplace,
install.
One file per server. Every server is a single .py — nothing to build, no
package tree to transfer. Three of the eight are standard library only, and
the standalone skills are prose (bar org-chart's one standard-library script).
Four settings, once, for the whole suite. EVAPYTHON,
EVADOCUMENTSDIR, EVATEMPLATESDIR and EVAKNOWLEDGEDIR are the
configuration. Each server works in its own sub-folder of those roots, named
after the plugin, so there is nothing per-plugin to fill in, no folder prompts
at install, and no folder flags to keep in step across a .mcp.json, a
claude mcp add line and eight plugin dialogs. /plugin install asks for
nothing: what is genuinely one server's own — a Confluence URL, an
embeddings endpoint — is an environment variable too, set in the same env
block. The matching skill comes with the server.
Confined by default. Every server that touches the filesystem is locked to
its sub-folders, and refuses to start unconfined rather than falling back to
"anywhere". The office plugins - word, powerpoint and excel - can
change a file you already have in their folder; the rest either read, or
write new files (Markdown, a downloaded attachment) into their own folders.
Shared systems are read-only until you say otherwise. confluence and
jira can write - create and edit pages, raise and move tickets - but each
ships with writing off. One variable (CONFLUENCEALLOWWRITE,
JIRAALLOWWRITE) turns it on; until then the write tools are not even
offered.
Secrets never hit the command line. Tokens and API keys are environment
variables only — argv is visible to other local users in process listings.
Listed badge (optional)
Copy this line into the README if you want the listing mark. Optional, and it does not affect being listed.
[![Listed on Anclum](https://anclum.com/badge/listed.svg)](https://anclum.com/en/s/mcnamee-claude-skills/)
Source
github · https://github.com/mcnamee/claude-skills
Browse by field
Open source
Traction
1 ★
added
today

More like this

All 1,671 →

Filed under

Every tag opens the full list for that filter. Counts are the live catalogue figures.

Browse all fields →

← Back to index