Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Use the Feishu/Lark plugin MCP tools for Feishu documents, spreadsheets, knowledge content, and other matching workspace operations.
.claude/skills/thinkinaixyz-feishu-tools/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-05 | ✗→✓ | ▲ Improved | 90% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -32% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 75% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -8% | 0% |
| case-17 | ✗→✓ | ▲ Improved | 31% | 0% |
This plugin is an MCP server tool surface exposed by DeepChat's Feishu plugin. Do not ask the user to classify the plugin as an MCP server, a CLI tool, or another plugin type. When a request is about Feishu/Lark content and matching tools are available, invoke the relevant tool directly.
${OWNER_PLUGIN_ID}.${PLUGIN_ROOT}.feishu-tools.documents.
workspace data.
matching tool by name or description.
feishu-tools MCP tools as the primary action surface for Feishu/Lark requests.the server supports.
of plugin it is.
workspace identifier when the target tool expects an id or token.
artifact or requested mutation is ambiguous or destructive.
Feishu preset may not include it and describe the gap.
Feishu plugin settings and verify App ID, App Secret, brand, and preset.
spreadsheets, tables, or bitable-like structures.
they are exposed by the current preset.
Tool availability depends on the current Feishu preset. The skill should guide tool choice, not invent unsupported tool names.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 2,091 | 8,073 | +286% | 1 | 1 | 0% | 367 | 1,288 | +251% | 0 | 0 | — |
case-02 | fail→fail | 3,032 | 13,884 | +358% | 1 | 1 | 0% | 567 | 2,156 | +280% | 0 | 0 | — |
case-03 | fail→fail | 3,004 | 6,836 | +128% | 1 | 1 | 0% | 445 | 1,182 | +166% | 0 | 0 | — |
case-12 | pass→pass | 2,869 | 2,727 | -5% | 1 | 1 | 0% | 505 | 1,043 | +107% | 0 | 0 | — |
case-04 | pass→pass | 2,724 | 2,872 | +5% | 1 | 1 | 0% | 440 | 1,023 | +133% | 0 | 0 | — |
case-05 | fail→pass | 3,170 | 2,507 | -21% | 1 | 1 | 0% | 472 | 896 | +90% | 0 | 0 | — |
case-06 | pass→pass | 10,578 | 2,781 | -74% | 1 | 1 | 0% | 1,618 | 1,003 | -38% | 0 | 0 | — |
case-07 | fail→pass | 12,546 | 5,414 | -57% | 1 | 1 | 0% | 2,167 | 1,473 | -32% | 0 | 0 | — |
case-13 | pass→fail | 6,419 | 4,541 | -29% | 1 | 1 | 0% | 1,174 | 1,280 | +9% | 0 | 0 | — |
case-08 | pass→pass | 11,400 | 5,213 | -54% | 1 | 1 | 0% | 2,091 | 1,382 | -34% | 0 | 0 | — |
case-09 | fail→fail | 3,770 | 3,033 | -20% | 1 | 1 | 0% | 682 | 983 | +44% | 0 | 0 | — |
case-10 | fail→pass | 6,238 | 8,301 | +33% | 1 | 1 | 0% | 1,182 | 2,065 | +75% | 0 | 0 | — |
case-11 | fail→pass | 7,011 | 3,446 | -51% | 1 | 1 | 0% | 1,179 | 1,089 | -8% | 0 | 0 | — |
case-14 | pass→pass | 6,986 | 3,943 | -44% | 1 | 1 | 0% | 1,325 | 1,227 | -7% | 0 | 0 | — |
case-15 | fail→fail | 2,761 | 8,705 | +215% | 1 | 1 | 0% | 360 | 1,343 | +273% | 0 | 0 | — |
case-16 | fail→fail | 2,788 | 10,537 | +278% | 1 | 1 | 0% | 437 | 1,093 | +150% | 0 | 0 | — |
case-17 | fail→pass | 6,470 | 5,551 | -14% | 1 | 1 | 0% | 1,102 | 1,439 | +31% | 0 | 0 | — |
case-22 | pass→pass | 2,221 | 2,449 | +10% | 1 | 1 | 0% | 335 | 975 | +191% | 0 | 0 | — |
case-18 | fail→pass | 6,165 | 4,297 | -30% | 1 | 1 | 0% | 1,098 | 1,264 | +15% | 0 | 0 | — |
case-19 | pass→pass | 10,601 | 6,444 | -39% | 1 | 1 | 0% | 1,888 | 1,404 | -26% | 0 | 0 | — |
case-20 | fail→pass | 9,110 | 2,816 | -69% | 1 | 1 | 0% | 1,460 | 965 | -34% | 0 | 0 | — |
case-21 | pass→pass | 3,018 | 4,297 | +42% | 1 | 1 | 0% | 415 | 1,271 | +206% | 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. 22 cases were attempted, and 17 counted toward the lift figure. The other 5 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 +27 percentage points is the difference between those two pass rates over the 17 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.