All authors

Claude Skills by rezzminator
github.com/rezzminator7 skills0 installs0 views
- AssetsUSER-ONLY Moves every open thread of the session to a NEW chat — the user types /handoff [--branch] [message]; never run it unasked. Writes a file, seeds a fresh chat from it: default reboots this pane into it; --branch starts it detached. Not for branching/forking this conversation → pfm chat branch.Votes: 0GitHub stars: 8
- TokensAttributes runtime token spend, heaviest first — Claude Code chats and sub-agents, Codex CLI threads with `--codex`, one flight's agents with `--flight <dir>`. Flags `--since 24h|3d`, `--project`, `--family`, `--session`, `--top N`, `--out FILE`, `--metrics-out FILE`. Triggers "token audit", "which agent burned the most", "what did the flight cost". Static context size → /context-meter.Votes: 0GitHub stars: 8
- CodeprobeInternal kit of collector — codeprobe.py copies code verbatim for numbered orders and prints their manifest. Not invoked directly — ask collector.Votes: 0GitHub stars: 8
- ChatMessages the tmux agent chats through `pfm chat` — `pfm chat inject {target} '{one line}'` sends a turn to a teammate's pane, `pfm whoami` gives your own address; read the delivery receipt. Use when a protocol says "ping", "inject", or "reply to {session}".Votes: 0GitHub stars: 8
- LegalIndex of curated legal/compliance playbooks — DPA (Art. 28), DPIA (Art. 35), breach response (Art. 33/34), privacy notices (Art. 13/14), vendor due diligence, red-team QC, NDA/risk triage, statute reading, pre-delivery self-check, skill-optimizer. Load for legal drafting, any DPA/DPIA/privacy/breach/consent doc, vendor vetting or legal QC. Formal compliance assessment → /officer.Votes: 0GitHub stars: 8
- Close OutUSER-ONLY Finishes a chat completely — /close-out: sweeps open threads, unfinished or unverified work, uncommitted changes, running agents; does what needs no ruling, asks the rest. Returns the closing report.Votes: 0GitHub stars: 8
- TranscriptDigests one agent session — "read that chat''s transcript", "what did the executor do": a Claude or Codex transcript by path, session id, agent id or chat name; `show|counts|types|locate|agents`, `--only --tool --grep -i --since --until --lines --first --last --results`, shape `--text --final --width --err-lines --out --root`. Returns one line per event or sub-agent. Spend → /tokens.Votes: 0GitHub stars: 8