Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Configure Lindy triggers, scheduling, multi-agent delegation, and automation. Use when setting up trigger-based workflows, scheduling agents, building multi-agent societies, or configuring agent delegation. Trigger with phrases like "lindy automation", "schedule lindy agent", "lindy workflow automation", "lindy delegation", "lindy multi-agent".
.claude/skills/jeremylongshore-lindy-core-workflow-b/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | -2% | 0% |
| case-01 | ✗→✓ | ▲ Improved | 15% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 24% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 22% | 0% |
| case-14 | ✗→✓ | ▲ Improved | 7% | 0% |
Configure trigger-based automation, scheduled agents, and multi-agent delegation ("Societies of Lindies"). Covers all trigger types, trigger filters, multiple triggers per agent, and the Agent Send Message delegation pattern.
lindy-core-workflow-a (agent creation)Trigger filters add conditional logic that determines WHICH events wake the agent. Filters are not case-sensitive and support AND/OR operators with condition groups.
Email trigger filter example:
Filter: sender contains "support" AND subject does not contain "auto-reply"Slack trigger filter example:
Filter: channel equals "#help-desk" AND message contains "urgent"One agent can respond to multiple trigger types feeding into the same workflow:
A single agent can run entirely different workflows based on trigger type:
Each trigger connects to its own action chain.
For recurring automation (daily reports, weekly digests):
Every weekday at 9:00 AMFor meeting-related automation:
-30 = trigger 30 minutes BEFORE event+5 = trigger 5 minutes AFTER event startsBuild modular agent systems where specialized agents collaborate:
Architecture:
[Lead Generator Lindy]
↓ Agent Send Message
[Outreach Lindy]
↓ Agent Send Message
[Meeting Scheduler Lindy]Setup:
Benefits:
Linked Actions create downstream execution paths after an action completes:
Multi-channel example:
Send Email → [Channel: Email Reply Received]
→ [Channel: Slack Message Received]
Each channel maintains its own conversation thread.For batch processing (e.g., process 50 leads from a spreadsheet):
| Trigger | Filter Options | Use Case | |---------|---------------|----------| | Webhook Received | Headers, body fields | External API integration | | Email Received | Sender, subject, label | Inbox automation | | Schedule | Time, recurrence | Recurring reports | | Chat Message | User, content | Interactive bot | | Slack Message | Channel, keyword, user | Team automation | | Agent Message | Sending agent | Multi-agent delegation | | Calendar Event | Event type, offset | Meeting automation | | Google Sheets Row | Sheet, column values | Data pipeline |
| Issue | Cause | Solution | |-------|-------|----------| | Trigger fires too often | No filter configured | Add trigger filters with AND/OR conditions | | Delegation fails | Target agent missing Agent trigger | Add Agent Message Received trigger to target | | Schedule not firing | Timezone mismatch | Verify timezone in agent settings | | Loop runs forever | No max cycles set | Always set Max Cycles limit | | Channel never activates | Reply sent to wrong thread | Verify thread context in channel config | | Context too large | Passing full history to delegate | Use selective context in Agent Send Message |
Proceed to lindy-common-errors for troubleshooting.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-02 | fail→pass | 20,566 | 13,892 | -32% | 1 | 1 | 0% | 2,550 | 2,491 | -2% | 0 | 0 | — |
case-01 | fail→pass | 20,574 | 15,257 | -26% | 1 | 1 | 0% | 2,542 | 2,931 | +15% | 0 | 0 | — |
case-03 | fail→pass | 23,871 | 15,517 | -35% | 1 | 1 | 0% | 2,538 | 3,152 | +24% | 0 | 0 | — |
case-04 | pass→pass | 10,857 | 7,263 | -33% | 1 | 1 | 0% | 914 | 1,621 | +77% | 0 | 0 | — |
case-05 | pass→pass | 16,333 | 8,081 | -51% | 1 | 1 | 0% | 1,655 | 1,759 | +6% | 0 | 0 | — |
case-06 | pass→pass | 16,576 | 9,130 | -45% | 1 | 1 | 0% | 1,838 | 2,000 | +9% | 0 | 0 | — |
case-07 | pass→pass | 15,669 | 8,987 | -43% | 1 | 1 | 0% | 1,706 | 1,983 | +16% | 0 | 0 | — |
case-08 | pass→pass | 11,298 | 7,960 | -30% | 1 | 1 | 0% | 1,076 | 1,732 | +61% | 0 | 0 | — |
case-09 | pass→pass | 10,480 | 7,630 | -27% | 1 | 1 | 0% | 984 | 1,685 | +71% | 0 | 0 | — |
case-10 | pass→pass | 15,918 | 8,288 | -48% | 1 | 1 | 0% | 1,695 | 1,865 | +10% | 0 | 0 | — |
case-11 | fail→pass | 21,277 | 15,560 | -27% | 1 | 1 | 0% | 2,183 | 2,674 | +22% | 0 | 0 | — |
case-12 | pass→pass | 11,084 | 7,569 | -32% | 1 | 1 | 0% | 970 | 1,558 | +61% | 0 | 0 | — |
case-13 | pass→pass | 8,812 | 8,272 | -6% | 1 | 1 | 0% | 1,209 | 1,829 | +51% | 0 | 0 | — |
case-14 | fail→pass | 14,505 | 10,696 | -26% | 1 | 1 | 0% | 1,888 | 2,025 | +7% | 0 | 0 | — |
case-15 | pass→pass | 9,180 | 3,898 | -58% | 1 | 1 | 0% | 1,248 | 1,809 | +45% | 0 | 0 | — |
case-16 | pass→pass | 11,430 | 2,776 | -76% | 1 | 1 | 0% | 845 | 1,594 | +89% | 0 | 0 | — |
case-17 | pass→pass | 13,743 | 9,662 | -30% | 1 | 1 | 0% | 1,825 | 1,855 | +2% | 0 | 0 | — |
case-18 | pass→pass | 15,467 | 3,562 | -77% | 1 | 1 | 0% | 1,476 | 1,829 | +24% | 0 | 0 | — |
case-19 | pass→pass | 10,036 | 7,330 | -27% | 1 | 1 | 0% | 819 | 1,573 | +92% | 0 | 0 | — |
case-20 | pass→pass | 14,670 | 7,757 | -47% | 1 | 1 | 0% | 1,356 | 1,733 | +28% | 0 | 0 | — |
case-21 | pass→pass | 24,055 | 17,751 | -26% | 1 | 1 | 0% | 2,957 | 3,056 | +3% | 0 | 0 | — |
case-22 | fail→pass | 29,374 | 16,992 | -42% | 1 | 1 | 0% | 2,949 | 2,996 | +2% | 0 | 0 | — |
case-23 | fail→fail | 23,114 | 17,572 | -24% | 1 | 1 | 0% | 2,619 | 3,315 | +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. 23 cases were attempted. The headline lift of +26 percentage points is the difference between those two pass rates over the 23 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.