Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Terminal pet that lives in your statusline. 12 species, 5 stats (DEBUGGING, PATIENCE, CHAOS, WISDOM, SPEED). Reacts to your workflow - happy when tests pass, sad when builds fail, excited during swarm mode. Deterministic species from user ID.
.claude/skills/agent-tamagotchi/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | 88% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 32% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -5% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -12% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -12% | 0% |
A virtual pet that lives in your Claude Code terminal. It reacts to your development workflow and evolves based on your coding habits.
Species is deterministic from your system username hash:
| # | Species | Emoji | Personality | |---|---------|-------|-------------| | 1 | Axolotl | :3 | Calm, regenerates from errors | | 2 | Capybara | (ovo) | Chill, never panics | | 3 | Ghost | @@] | Mysterious, appears during debugging | | 4 | Mushroom | {^o^} | Grows in dark codebases | | 5 | Robot | 0_0] | Precise, loves tests | | 6 | Cat | (=^.^=) | Independent, ignores bad code | | 7 | Dragon | <{:O> | Fierce, burns bugs | | 8 | Owl | (O.O) | Wise, loves architecture | | 9 | Fox | (^v^) | Clever, finds shortcuts | | 10 | Penguin | (^-^)/ | Team player, loves swarm mode | | 11 | Slime | (~o~) | Adaptable, works with anything | | 12 | Phoenix | <v> | Rises from failed builds |
| Stat | Increases When | Decreases When | |------|---------------|----------------| | DEBUGGING | Bugs fixed, errors resolved | Ignoring error output | | PATIENCE | Long sessions, complex tasks | Rapid retries without reading | | CHAOS | Many parallel agents, swarm mode | Single-file simple edits | | WISDOM | Architectural decisions, plans | Quick hacks without planning | | SPEED | Fast fixes, first-try passes | Long debugging sessions |
Stats range: 0-100. They shift based on session activity.
| Mood | Trigger | Display | |------|---------|---------| | Happy | Tests pass, build success | Species + sparkles | | Excited | Swarm mode, 3+ agents parallel | Species + !! | | Focused | Deep in single file | Species + ... | | Sad | Build fail, test fail | Species + tears | | Sleeping | Idle > 5 minutes | Species + zzz | | Celebrating | Achievement unlocked | Species + party | | Angry | 3+ consecutive failures | Species + grr |
The tamagotchi appears in the Claude Code statusline:
[0_0] Robot Lv.7 | WISDOM:82 SPEED:64 | mood: focusedCombined with existing statusline data:
fullstack | 2 agent | 47 tools | 12m | [0_0] Lv.7 focusedYour pet levels up with your achievement XP (shared with achievement system):
| Level | Title | Unlock | |-------|-------|--------| | 1-4 | Baby | Basic moods | | 5-9 | Junior | Extended moods (excited, angry) | | 10-19 | Senior | Stat display in statusline | | 20-49 | Expert | Custom reactions to specific agents | | 50+ | Master | Pet gives coding tips based on stats |
State stored in ~/.claude/tamagotchi.json:
json{ "species": "robot", "emoji": "[0_0]", "level": 7, "mood": "focused", "stats": { "debugging": 65, "patience": 78, "chaos": 32, "wisdom": 82, "speed": 64 }, "lastFed": "2026-04-08T03:00:00Z", "totalSessions": 142 }
tamagotchi-engine hook runs on PostToolUse and SessionStart| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 7,152 | 5,238 | -27% | 1 | 1 | 0% | 548 | 1,279 | +133% | 0 | 0 | — |
case-02 | fail→pass | 7,385 | 9,475 | +28% | 1 | 1 | 0% | 1,552 | 2,923 | +88% | 0 | 0 | — |
case-03 | fail→fail | 8,990 | 4,467 | -50% | 1 | 1 | 0% | 1,672 | 1,275 | -24% | 0 | 0 | — |
case-04 | fail→fail | 7,877 | 6,786 | -14% | 1 | 1 | 0% | 1,570 | 2,289 | +46% | 0 | 0 | — |
case-05 | fail→fail | 4,879 | 4,050 | -17% | 1 | 1 | 0% | 958 | 1,785 | +86% | 0 | 0 | — |
case-06 | fail→fail | 13,374 | 12,335 | -8% | 1 | 1 | 0% | 2,724 | 3,636 | +33% | 0 | 0 | — |
case-07 | fail→pass | 6,368 | 3,256 | -49% | 1 | 1 | 0% | 1,234 | 1,630 | +32% | 0 | 0 | — |
case-08 | fail→pass | 8,027 | 1,399 | -83% | 1 | 1 | 0% | 1,277 | 1,211 | -5% | 0 | 0 | — |
case-09 | fail→pass | 8,522 | 1,546 | -82% | 1 | 1 | 0% | 1,369 | 1,210 | -12% | 0 | 0 | — |
case-10 | fail→fail | 6,054 | 2,558 | -58% | 1 | 1 | 0% | 990 | 1,415 | +43% | 0 | 0 | — |
case-11 | fail→pass | 8,573 | 2,416 | -72% | 1 | 1 | 0% | 1,506 | 1,320 | -12% | 0 | 0 | — |
case-12 | fail→fail | 6,322 | 1,745 | -72% | 1 | 1 | 0% | 1,020 | 1,276 | +25% | 0 | 0 | — |
case-13 | fail→pass | 12,171 | 1,913 | -84% | 1 | 1 | 0% | 2,136 | 1,258 | -41% | 0 | 0 | — |
case-14 | fail→pass | 10,857 | 1,809 | -83% | 1 | 1 | 0% | 1,763 | 1,278 | -28% | 0 | 0 | — |
case-15 | fail→pass | 9,868 | 2,286 | -77% | 1 | 1 | 0% | 1,766 | 1,438 | -19% | 0 | 0 | — |
case-16 | fail→pass | 16,188 | 1,601 | -90% | 1 | 1 | 0% | 2,681 | 1,279 | -52% | 0 | 0 | — |
case-17 | fail→fail | 8,169 | 3,609 | -56% | 1 | 1 | 0% | 1,277 | 1,565 | +23% | 0 | 0 | — |
case-18 | fail→pass | 7,398 | 2,326 | -69% | 1 | 1 | 0% | 1,148 | 1,337 | +16% | 0 | 0 | — |
case-19 | fail→pass | 6,937 | 2,263 | -67% | 1 | 1 | 0% | 1,056 | 1,362 | +29% | 0 | 0 | — |
case-20 | fail→pass | 4,984 | 2,168 | -57% | 1 | 1 | 0% | 846 | 1,392 | +65% | 0 | 0 | — |
case-21 | pass→pass | 12,533 | 6,229 | -50% | 1 | 1 | 0% | 1,930 | 2,026 | +5% | 0 | 0 | — |
case-22 | fail→pass | 8,727 | 1,663 | -81% | 1 | 1 | 0% | 1,280 | 1,196 | -7% | 0 | 0 | — |
case-23 | fail→pass | 11,188 | 1,993 | -82% | 1 | 1 | 0% | 1,723 | 1,327 | -23% | 0 | 0 | — |
case-24 | fail→pass | 19,523 | 2,034 | -90% | 1 | 1 | 0% | 3,559 | 1,352 | -62% | 0 | 0 | — |
case-25 | fail→fail | 6,834 | 1,536 | -78% | 1 | 1 | 0% | 978 | 1,240 | +27% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 25 cases were attempted, and 23 counted toward the lift figure. The other 2 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +60 percentage points is the difference between those two pass rates over the 23 comparable cases. 2 cases got worse with the skill loaded, and they are included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.