Fullstop000

Use when verifying an ignis change by running it like a real user before shipping — especially TUI/visual changes, where you must SEE the rendered output. The /ship step 3 routes here. Triggers include "dogfood", "does it actually work", "see the TUI", "screenshot the diff/picker".

Ship ignis — land a PR on master, optionally bump+tag for cross-platform release. Triggers: ship, release, cut a release, deploy ignis.

Use this skill when the user asks about token usage, API costs, spending, or session statistics for Claude Code, Codex CLI, Kimi Code, or Gemini CLI. This skill runs the `toll` CLI to aggregate token counts and estimated USD costs from local session logs stored in ~/.claude, ~/.codex, ~/.kimi, and ~/.gemini. Activate when the user says things like "how many tokens have I used", "how much have I spent on Claude", "show my usage", "token stats", "daily usage breakdown", or "export my usage as CSV".

Google Gemini CLI wrapper for independent second-opinion code review. Three modes: (1) Review — diff-based review with [P1]/[P2] pass/fail gate, (2) Challenge — adversarial audit for edge cases, security holes, race conditions, (3) Consult — free-form Q&A about the codebase with session continuity. Use when the user asks for: "gemini review", "gemini challenge", "ask gemini", "consult gemini", "second opinion from gemini", or any request to get an independent AI review of code changes using the Gemini CLI.