Lindy Debug BundleSAFE

Model-agnostic agent-skills platform with a harness-free canonical layer, verified adapters, and the ccpi package manager. Explore at tonsofskills.com.

The instruction file, verbatim from the audited commit — this is the text the model reads, and the surface the audit's instruction layer examines. Quoted here so you can judge it without cloning anything.

---
name: lindy-debug-bundle
description: 'Comprehensive debugging toolkit for Lindy AI agents.

  Use when investigating complex agent failures, collecting diagnostics,

  or preparing support tickets for Lindy support.

  Trigger with phrases like "lindy debug", "lindy diagnostics",

  "lindy support bundle", "investigate lindy issue".

  '
allowed-tools: Read, Write, Edit, Bash(curl:*), Grep
version: 1.20.0
license: MIT
author: Jeremy Longshore <[email protected]>
tags:
- saas
- lindy
- debugging
compatibility: Designed for Claude Code
---
# Lindy Debug Bundle

## Current State

!`node --version 2>/dev/null || echo 'Node.js not installed'`
!`python3 --version 2>/dev/null || echo 'Python not installed'`
!`curl --version 2>/dev/null | head -1 || echo 'curl not installed'`

## Overview

Systematic diagnostics for Lindy AI agent issues. Collects environment info,
tests API connectivity, reviews agent task history, and generates a support
bundle for Lindy's support team.

## Prerequisites

- Access to Lindy dashboard (https://app.lindy.ai)
- curl installed for API testing
- Agent ID and webhook URLs available

## Instructions

### Step 1: Collect Environment Info

```bash
# Local environment diagnostics
echo "=== Local Environment ==="
echo "Node: $(node --version 2>/dev/null || echo 'N/A')"
echo "Python: $(python3 --version 2>/dev/null || echo 'N/A')"
echo "OS: $(uname -srm)"
echo "Date: $(date -u '+%Y-%m-%d %H:%M:%S UTC')"
echo "LINDY_API_KEY set: $([ -n "$LINDY_API_KEY" ] && echo 'yes' || echo 'NO')"
echo "LINDY_WEBHOOK_SECRET set: $([ -n "$LINDY_WEBHOOK_SECRET" ] && echo 'yes' || echo 'NO')"
```

### Step 2: Test Webhook Connectivity

```bash
# Test webhook trigger endpoint
echo "=== Webhook Connectivity ==="
WEBHOOK_URL="${LINDY_WEBHOOK_URL:-https://public.lindy.ai/api/v1/webhooks/YOUR_ID}"

# Test without auth (expect 401)
echo "Without auth (expect 401):"
curl -s -o /dev/null -w "HTTP %{http_code} in %{time_total}s\n" \
  -X POST "$WEBHOOK_URL" \
  -H "Content-Type: application/json" \
  -d '{"test": true}'

# Test with auth (expect 200)
echo "With auth (expect 200):"
curl -s -o /dev/null -w "HTTP %{http_code} in %{time_total}s\n" \
  -X POST "$WEBHOOK_URL" \
  -H "Authorization: Bearer $LINDY_WEBHOOK_SECRET" \
  -H "Content-Type: application/json" \
  -d '{"test": true, "debug": "bundle-test"}'
```

### Step 3: Review Agent Task History

In the Lindy dashboard:

1. Navigate to the failing agent
2. Open the **Tasks** tab
3. Filter by **Failed** status
4. For each failed task:
   - Note the timestamp
   - Click to expand step-by-step execution
   - Identify the failing step (marked red)
   - Copy the error message and input/output data
5. Look for patterns: same step failing? same time of day? same input type?

### Step 4: Check Integration Health

```bash
# Test outbound connectivity to common Lindy integration targets
echo "=== Integration Targets ==="
for url in \
  "https://public.lindy.ai" \
  "https://slack.com/api/auth.test" \
  "https://www.googleapis.com/gmail/v1/users/me/profile" \
  "https://api.notion.com/v1/users/me"
do
  status=$(curl -s -o /dev/null -w "%{http_code}" --max-time 5 "$url" 2>/dev/null)
  echo "$url -> HTTP $status"
done
```

### Step 5: Diagnose Specific Failure Types

**Trigger not firing**:

- Verify agent status: active (not paused)
- Check trigger filter conditions
- For webhooks: test URL with curl
- For email: re-authorize Gmail/Outlook
- For schedule: verify timezone settings

**Action failing**:

- Check integration authorization (re-auth if token expired)
- Verify field references: `{{step_name.field}}` syntax correct
- Test the target service independently
- Check if action is a Premium Action (requires Pro plan)

**Agent step looping**:

- Review exit conditions — are they achievable?
- Check credit consumption (rapid drain = looping)
- Reduce available skills to 2-4 focused ones
- Add a fallback exit condition

**High credit consumption**:

- Review model selection: Gemini Flash (cheap) vs GPT-4 (expensive)
- Check for unnecessary agent steps (use deterministic actions instead)
- Review loop configurations for unbounded max cycles

### Step 6: Generate Support Bundle

Compile the following for a support ticket to [email protected]:

```markdown
## Lindy Support Bundle

**Account**: [your email]
**Agent Name**: [agent name]
**Agent URL**: agents/[agent-id]
**Issue Start**: [date/time UTC]
**Frequency**: [every time / intermittent / once]

### Environment
- Browser: [Chrome/Firefox/Safari version]
- Plan: [Free/Pro/Business/Enterprise]
- Credit Balance: [remaining credits]

### Reproduction Steps
1. [step 1]
2. [step 2]

### Failed Task IDs
- [task-id-1] at [timestamp]
- [task-id-2] at [timestamp]

### Error Messages
[Copy exact error text from task detail view]

### What I Tried
- [attempt 1]
- [attempt 2]
```

## Output

Deliver a time-bounded support bundle containing the affected agent and task
IDs, reproduction steps, sanitized error messages, environment facts, and the
result of the webhook and integration checks. Remove credentials, personal
data, and full webhook URLs before sharing the bundle outside the workspace.

## Examples

If tasks begin failing after an integration reconnect, capture one failed task
ID, the failing action name, the HTTP status from the target, and the time the
connection was reauthorized. A bundle saying “agent failed” without a task ID,
timestamp, and reproducible trigger is insufficient for escalation.

## Diagnostic Decision Tree

```
Agent not working?
├── No task created → Check trigger configuration
│   ├── Webhook? → Test with curl (Step 2)
│   ├── Email? → Re-authorize + check filters
│   └── Schedule? → Check timezone + credit balance
├── Task created but failed → Check task detail view
│   ├── Trigger step failed → Auth/connectivity issue
│   ├── Action step failed → Integration auth expired
│   ├── Condition step failed → Ambiguous condition prompt
│   └── Agent step looping → Exit conditions un

SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.

What the source does

Filesystem

none-observed

Network

none-observed

Shell

none-observed

Dependencies

pinned

Secrets in source

none-found

Findings (0)

No findings outside the package's declared scope.

Gates applied: no_behavioural_pass.