Loading...
When an agent tool call fails, this skill analyzes the error (rate limit, auth expired, invalid input, network timeout, etc.), maps it to your configured recovery strategy (retry with backoff, switch provider, escalate to human, skip and continue), and returns structured instructions the agent can act on. Reduces cascading failures in multi-step workflows.
curl -X POST https://jarvissdk.com/api/v1/skills/agent-error-recovery/execute \
-H "x-api-key: jsk_your_key" \
-H "Content-Type: application/json" \
-d '{
"input": "your data here"
}'Install this skill and let your AI agent run it automatically. Free tier includes 1,000 executions/month.
Encodes your incident response process — severity classification, escalation paths, runbook selection, status page updates, and post-mortem templates.
Breaks complex instructions into atomic, executable subtasks for AI agents. Encodes your decomposition patterns, dependency ordering, and validation criteria.
Persists and retrieves agent memory across sessions — facts, decisions, user preferences, and context. Encodes your memory taxonomy and retention policies.
Tracks API rate limit headers across multiple services, alerts when approaching limits, and suggests request throttling strategies.