Imported from shortcuts/dotfiles (.claude/skills/radin-stats/SKILL.md). Install upstream with npx skills add shortcuts/dotfiles --skill radin-stats. Copyright stays with the author.

Stats Roundup

Print each installed tool's own stats or gain command, back to back. Do not compute a merged total: the numbers use incompatible units -- real per-session tokens, static benchmark medians, and a counted per-repo ledger. A sum would misrepresent all three. Display each output as-is.

Step 1: Invoke each available source

Run whichever sources are installed, checking with command -v or a skill lookup. Skip a missing source without printing anything, and never treat it as an error.

• /caveman:caveman-stats -- real per-session token usage and savings, read from the session log itself.
• /ponytail:ponytail-gain -- ponytail's published benchmark scoreboard (medians across 5 tasks, 3 models). Not this session, and not this repo.
• /ponytail:ponytail-debt -- ponytail's real per-repo ledger of deferred shortcuts, if the repo has one.
• rtk gain -- rtk's real token-savings ledger (command -v rtk; pass -p to scope it to the current project).
• headroom savings -- headroom's real compression ledger (command -v headroom). "No savings recorded yet" is a normal empty result: nothing has been routed through its proxy or MCP tool.

Any other installed tool with its own stats or gain command belongs in this list too. Add it here rather than building a separate skill.

If a source finds nothing to report, treat that as a normal empty result: one quick check, then move on. Do not keep searching for entries that are not there. This output is informative only, so it does not need to be exact.

Step 2: Display, don't merge

Print each tool's output under its own heading, in the order above. Label which numbers are really measured (caveman-stats, ponytail-debt, rtk gain, headroom savings) and which come from a fixed benchmark (ponytail-gain), so the user does not mistake one for the other. Print no combined total row.