Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Diagnose and fix MCP configuration/runtime issues across supported integrations using agents CLI checks.
.claude/skills/bilal140202-mcp-troubleshooting/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | -73% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -69% | 0% |
| case-01 | ✗→✓ | ▲ Improved | -51% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -70% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -86% | 0% |
When MCP servers are failing or not visible:
agents status --fast to inspect configured vs enabled integrations.agents mcp test for schema/config validation.agents mcp test --runtime to check integration CLI/runtime connectivity.agents doctor for environment and binary diagnostics.agents sync and verify with agents sync --check.Troubleshooting priorities:
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-02 | fail→pass | 16,777 | 2,921 | -83% | 1 | 1 | 0% | 2,561 | 691 | -73% | 0 | 0 | — |
case-03 | fail→pass | 15,682 | 4,091 | -74% | 1 | 1 | 0% | 2,587 | 801 | -69% | 0 | 0 | — |
case-01 | fail→pass | 14,731 | 5,942 | -60% | 1 | 1 | 0% | 2,492 | 1,209 | -51% | 0 | 0 | — |
case-04 | fail→pass | 11,397 | 2,225 | -80% | 1 | 1 | 0% | 1,683 | 500 | -70% | 0 | 0 | — |
case-05 | fail→pass | 16,105 | 1,957 | -88% | 1 | 1 | 0% | 2,566 | 365 | -86% | 0 | 0 | — |
case-06 | fail→pass | 10,297 | 2,407 | -77% | 1 | 1 | 0% | 1,554 | 519 | -67% | 0 | 0 | — |
case-07 | fail→pass | 12,866 | 3,330 | -74% | 1 | 1 | 0% | 1,925 | 586 | -70% | 0 | 0 | — |
case-08 | fail→pass | 9,308 | 2,469 | -73% | 1 | 1 | 0% | 1,545 | 529 | -66% | 0 | 0 | — |
case-09 | pass→pass | 6,782 | 2,599 | -62% | 1 | 1 | 0% | 991 | 476 | -52% | 0 | 0 | — |
case-10 | fail→pass | 12,462 | 5,020 | -60% | 1 | 1 | 0% | 2,008 | 896 | -55% | 0 | 0 | — |
case-11 | pass→pass | 4,865 | 3,396 | -30% | 1 | 1 | 0% | 764 | 650 | -15% | 0 | 0 | — |
case-12 | fail→pass | 19,332 | 2,630 | -86% | 1 | 1 | 0% | 1,290 | 529 | -59% | 0 | 0 | — |
case-13 | pass→pass | 9,467 | 2,831 | -70% | 1 | 1 | 0% | 1,333 | 578 | -57% | 0 | 0 | — |
case-14 | fail→pass | 12,939 | 3,276 | -75% | 1 | 1 | 0% | 2,151 | 724 | -66% | 0 | 0 | — |
case-15 | fail→pass | 4,652 | 2,363 | -49% | 1 | 1 | 0% | 844 | 493 | -42% | 0 | 0 | — |
case-16 | fail→pass | 7,493 | 2,525 | -66% | 1 | 1 | 0% | 1,171 | 514 | -56% | 0 | 0 | — |
case-17 | pass→pass | 7,306 | 2,978 | -59% | 1 | 1 | 0% | 1,167 | 279 | -76% | 0 | 0 | — |
case-18 | fail→pass | 7,443 | 2,956 | -60% | 1 | 1 | 0% | 1,070 | 661 | -38% | 0 | 0 | — |
case-19 | fail→pass | 11,429 | 2,923 | -74% | 1 | 1 | 0% | 2,036 | 656 | -68% | 0 | 0 | — |
case-20 | pass→pass | 11,644 | 6,744 | -42% | 1 | 1 | 0% | 2,374 | 1,428 | -40% | 0 | 0 | — |
case-21 | pass→pass | 6,504 | 5,311 | -18% | 1 | 1 | 0% | 1,202 | 892 | -26% | 0 | 0 | — |
case-22 | pass→pass | 6,673 | 4,910 | -26% | 1 | 1 | 0% | 1,321 | 1,052 | -20% | 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. The headline lift of +68 percentage points is the difference between those two pass rates over the 22 comparable cases.
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.