Watches subagents in long, fan-out-heavy sessions: a [GOV] band while agents run (count, peak, tokens) and a toast when too many run at once or an agent or the session goes over its token budget. Warn-only by default; /governor caps on asks before over-limit spawns and stops over-budget agents, /governor route picks a model per agent type.
Picks the model per prompt with local rules: quick questions to Haiku, writing and explaining to Sonnet, building and fixing to Opus. The whole conversation always goes along; it switches only when the other model's prompt-cache re-read costs little.
Holds risky shell commands (rm -rf, git reset --hard, git clean, force push, migrations) and shows what they would change, with Proceed and Cancel buttons.
A Claude Code mod that keeps the one-hour prompt cache warm during breaks and stops a cold send once with its estimated rewrite cost. Warming uses model requests and costs tokens. Requires Claude Code 2.1.287+.
Games above the Claude Code prompt (snake, Tetris, 2048, Minesweeper, Flappy, Pong, typing test, Space Invaders, Doom) plus a pet that grows as Claude tests, commits and edits. Clones of the genre, not affiliated with or endorsed by Tetris Holding, Taito, Atari or id Software. Needs function hooks (early access) and an interactive terminal.
Messenger-style chat for Claude Code: your prompts on the right in a rival color, Claude on the left, only running or failing tools stand out. 441 themes. (Vendored from github.com/angeldelbiondo/claude-chat-bubbles @ 5597448, MIT. Modified for my-claude-mods: the footer keeps the row beneath, so plan-progress's button stays; other mods' panes get the backdrop without side padding; an empty...
At every compaction, saves a handoff note (goal, open tasks, files, recent errors) to .claude/handoff/ and asks the summary to keep those facts; a later fresh session offers to resume from it (only a note this mod wrote, unchanged since). /handoff saves one now. No model calls.
An IDE-style file tree for Claude Code that shows what Claude is doing and where in files
Flightdeck: a live agent dashboard for Claude Code. Main model vitals, an on-call architect, every permission check, subagent cards and swimlanes, a turn receipt and a session log, all from real session events
/changes pane: branch, ahead/behind, staged, unstaged, untracked and conflicted files with +/- line counts, recent commits. Read-only git, refreshed as Claude works.
Asks before Claude runs or writes code that drives hardware: nidaqmx DO/AO writes, pymodbus coil/register writes, pyserial writes, MQTT publishes to printer topics.
Mission-control skin: [SYS]/[TOOL]/[EDIT] tool rows with per-tool icons and one status block above the prompt (under flightdeck's session log while its pane is open): [SYS] repo, branch, model, hw-guard; [CTL] model and effort buttons; [CTX] context fill with a /ctx pane by tool. /clean for a checklist view.
At the end of a turn Claude offers 2-4 follow-ups as buttons above the prompt, each showing the whole prompt it sends; a click sends that one. Gone as soon as you type something else.
Live plan progress bars above the Claude Code prompt: stages, steps, a pixel fill and soft sounds for decision, error and done
/improve : a pane asks 2-4 clickable questions about what you want, then writes a fuller prompt you can run, edit or copy. Two small model calls per upgrade, only when you ask.
/qa pane: every question you asked this session with the start of Claude's answer; click to jump the transcript to the answer or the question.
Redacts secrets and PII from every row Claude Code stores and from the CLAUDE.md and memory context, before the model reads it and before the transcript keeps it. Restores a value only into a file that already holds it. No network, no processes, no file writes.
When you correct Claude ("no", "don't", "I wish", "remember: ..."), a band offers to save it as a rule in CLAUDE.md or a skill. Local phrase matching, no model calls.
claude-reflect as a Claude Code mod: a model check on each short prompt finds reusable corrections (any language); a band above the prompt saves them to CLAUDE.md on one press
Replay Theater: step through the file edits Claude made in the last turn, one diff at a time. (Vendored from anthropic claude-code-playground @ 569c528, patched so its band row stacks above other mods' rows, and sits under flightdeck's session log while that pane is open.)
Warns (never blocks) when Claude or a subagent keeps retrying the same failing command: a toast and a row above the prompt after 3 identical failures in a row.
Progress bar above the prompt for /savvy-flow, plus an agents panel (/agents-info): running, finished and planned subagents with model, task progress, context, cost and time.
Asks before Claude edits a file that another tool (Gemini, an editor, another agent) changed in the last few minutes, through Edit/Write, a shell command that writes, or an MCP write tool. A Cancel keeps asking until you say Edit anyway or the change ages out.
/report: tasks Claude finished today, this week, month or year, the time it worked, what the work was, which models did it, the API-equivalent cost, your streak, and an estimate of time saved at a rate you set. Counted locally, no model calls.