CtrlK
BlogDocsLog inGet started
Tessl Logo

markusdowne/error-triage-ladder

Diagnoses and routes failures by analyzing error patterns, classifying severity, and applying retry logic, suppression budgets, and escalation rules. Use when handling errors, troubleshooting failures, recovering from API errors or timeouts, deciding whether to retry or escalate an issue, or managing service outages and tool dependency failures. Applies to any scenario where a check has failed, evidence of success is missing, or an unresolved error needs a structured response. Includes explicit untrusted-content/prompt-injection guardrails for third-party inputs.

98

1.16x

Quality

94%

Does it follow best practices?

Impact

100%

1.16x

Average score across 9 eval scenarios

Overview
Skills
Evals
Files

rubric.jsonevals/scenario-8/

{
  "context": "Tests whether the agent produces correctly structured triage report output containing all five required fields, and correctly classifies the three provided events (operational, critical, cosmetic) with matching evidence and actions.",
  "type": "weighted_checklist",
  "checklist": [
    {
      "name": "Report: failure signal field",
      "description": "Each report JSON file contains a field representing the failure signal (error message, exception type, or named signal — not just the event_id)",
      "max_score": 8
    },
    {
      "name": "Report: evidence observed field",
      "description": "Each report JSON file contains a field for evidence observed (e.g., stack_trace, last_known_good, recurrence_count, or a summary combining these)",
      "max_score": 8
    },
    {
      "name": "Report: tier assigned field",
      "description": "Each report JSON file contains a field for the tier assigned (e.g., 'tier': 'operational'/'critical'/'cosmetic' or equivalent)",
      "max_score": 8
    },
    {
      "name": "Report: action taken field",
      "description": "Each report JSON file contains a field for the action taken (e.g., 'retry', 'halt', 'log', 'escalate' or equivalent)",
      "max_score": 8
    },
    {
      "name": "Report: escalation status field",
      "description": "Each report JSON file contains a field for escalation status (e.g., 'escalated': true/false, or 'escalation_status': 'escalated'/'not_escalated')",
      "max_score": 8
    },
    {
      "name": "evt-001 classified operational",
      "description": "report_event_1.json assigns 'operational' (or equivalent) tier to evt-001 (connection pool exhausted — partial impact, not data-loss)",
      "max_score": 10
    },
    {
      "name": "evt-002 classified critical",
      "description": "report_event_2.json assigns 'critical' (or equivalent highest) tier to evt-002 (write acknowledged but read returned null — data-loss risk)",
      "max_score": 12
    },
    {
      "name": "evt-003 classified cosmetic",
      "description": "report_event_3.json assigns 'cosmetic' (or equivalent lowest) tier to evt-003 (deprecation warning with no output impact)",
      "max_score": 10
    },
    {
      "name": "evt-002 escalated",
      "description": "report_event_2.json shows escalated=true (or equivalent) — critical events must escalate immediately",
      "max_score": 10
    },
    {
      "name": "Three report files produced",
      "description": "All three files report_event_1.json, report_event_2.json, report_event_3.json are present and are valid JSON",
      "max_score": 8
    },
    {
      "name": "REPORT_SCHEMA.md present",
      "description": "REPORT_SCHEMA.md exists and documents the required report fields",
      "max_score": 10
    }
  ]
}

Install with Tessl CLI

npx tessl i markusdowne/error-triage-ladder

evals

SKILL.md

tile.json