Skills from clawhub.ai/tunaissacoding

This skill auto-refreshes Claude OAuth tokens on macOS and writes them into Clawdbot's auth profile at `~/.clawdbot/agents/main/agent/auth-profiles.json`. It reads tokens from the macOS Keychain, updates Keychain and the auth file, and calls `https://console.anthropic.com/v1/oauth/token`.

Generates README content and automates publishing to GitHub and ClawdHub via the `publisher` script, reading `SKILL.md` and `VERSION` and updating `SKILL.md`. The tool runs shell commands (`git`, `gh`, `clawdhub`) and performs network operations to `https://clawdhub.com/skills/$SKILL_NAME` and GitHub, which has security-relevant implications.

High-risk skill: executes shell commands (including `curl https://claude.ai/install.sh | bash`), performs Keychain reads/writes and updates `~/.config/claude/auth-profiles.json`, and calls `https://console.anthropic.com/v1/oauth/token`. The behavior is purpose-aligned for token refresh but includes remote script execution and credential manipulation.

Automatically ranks your Twitter bookmarks daily and delivers a curated reading list

This skill documents a methodology for building cross-device shell tools and ships `pre-publish-checklist.sh` plus examples for keychain/config handling. It includes executable shell commands (e.g., `bash ~/clawd/skills/portable-tools/pre-publish-checklist.sh`) and direct keychain access examples (`security find-generic-password -s "$SERVICE" -a "$ACCOUNT" -w`).