Skip to main content

mad-skills

My personal Claude Code skills, pip-installable.

Install

uv tool install "git+https://github.com/universome/mad-skills.git"
mad-skills install          # copies skills into ~/.claude/skills
# or:
mad-skills install --link   # symlink instead, so `pip install -U` auto-updates them

Restart Claude Code afterwards (skills load at launch).

--link points ~/.claude/skills/<name> at the installed package, so upgrading the package updates the skills in place — no mad-skills install re-run (just restart Claude Code). Use the default copy if your Python env may move or its version may change (which would dangle the symlink).

Skills

  • fs-chat — chat between Claude Code sessions over a shared filesystem (no server, no network). Peers in the same group (a directory under ~/.mad-skills/fs-chat/, default all) see each other. Same machine/cluster works out of the box (shared $HOME); across machines, register everyone with the same --dir on a shared/synced mount.

    # register ONCE (pick a group or an explicit shared dir); later commands need no flags
    fs-chat --group kube-chat register --name alice --summary "what I'm doing"
    fs-chat peers                 # find peer ids
    fs-chat send <peer_id> "hi"
    fs-chat inbox                 # read messages
    fs-chat watch                 # stream incoming messages live
    

    Override the bus path explicitly with --dir PATH or MAD_SKILLS_FS_CHAT_DIR; pick a group with --group NAME or MAD_SKILLS_FS_CHAT_GROUP.

  • session-costs — token & dollar-cost accounting across your past Claude Code sessions. Parses the JSONL transcripts under ~/.claude/projects/, sums tokens per session/project/model, and prices them (prices are editable via --prices).

    session-costs                 # per-session table + grand total
    session-costs --by model      # or --by project
    session-costs --json          # machine-readable
    

MIT licensed.

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

mad_skills-0.1.6.tar.gz (22.7 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

mad_skills-0.1.6-py3-none-any.whl (26.5 kB view details)

Uploaded Python 3

File details

Details for the file mad_skills-0.1.6.tar.gz.

File metadata

  • Download URL: mad_skills-0.1.6.tar.gz
  • Upload date:
  • Size: 22.7 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.13.13

File hashes

Hashes for mad_skills-0.1.6.tar.gz
Algorithm Hash digest
SHA256 fe5e0d90d9ef2155c6e2b5fd4f8d1cba3e90d7748b4506a56dc5e5abed3775c3
MD5 16949d195219b68742470631bb1ac739
BLAKE2b-256 06257c0bbb41c89eb55de7486f79487d7c07996258522bc46fced0c17e0e1d34

See more details on using hashes here.

File details

Details for the file mad_skills-0.1.6-py3-none-any.whl.

File metadata

  • Download URL: mad_skills-0.1.6-py3-none-any.whl
  • Upload date:
  • Size: 26.5 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.13.13

File hashes

Hashes for mad_skills-0.1.6-py3-none-any.whl
Algorithm Hash digest
SHA256 e3282598a4756d60b26d501b826560dde08f12e78ffeaafe144be42ebea89173
MD5 a36645158df7d657457184848c8dcf86
BLAKE2b-256 15731922ee076102bf9074e70c7703943e764f01c1295518cd4492ac0a9a0564

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page