triage

Run a single-environment triage cycle ingesting watchdog entries and deduplicating errors.

Updated Feb 28, 2026

Search Tags:#observability #deduplication #fingerprinting #error triage #slack notifications #ticket automation #kanban promotion

Ready to run

npx skills add https://github.com/cosmicdreams/claude-plugins --skill triage-cosmicdreams

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: triage
Source: https://github.com/cosmicdreams/claude-plugins/tree/main/drover/skills/triage
Command: npx skills add https://github.com/cosmicdreams/claude-plugins --skill triage-cosmicdreams

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Triage cycles often produce noisy, duplicate error reports and slow routing, making it hard to reliably turn raw watchdog noise into actionable tickets.

Core Features & Use Cases

Single-environment triage cycle: runs one complete pass for a named environment (watchdog gather → fingerprint → dedupe → promotion → notifications).
Fingerprint-based deduplication: computes stable hashes from structured record fields so the same error yields the same ticket fingerprint across entry points.
Trust-aware noise filtering & promotion: applies low-trust noise filters, then promotes tickets to lane-ready using thresholds and escalation signals (including cross-environment and velocity boosts).

Quick Start

Ask drover to run one triage pass for your environment by providing its configured name, then review the New/Augmented/Promoted counts in the cycle summary.

Frequently Asked Questions about triage

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I deduplicate watchdog error reports in a monitoring workflow?▼

Error deduplication uses fingerprint-based hashing of structured record fields to compute stable hashes, ensuring identical errors yield the same ticket fingerprint across entry points. This eliminates noisy, duplicate error reports during triage cycles.

How do I automate error triage and ticket promotion for a specific environment?▼

Run a single-environment triage cycle that ingests watchdog entries, applies deduplication, evaluates promotion rules using trust thresholds and escalation signals, then sends notifications. One pass processes the complete flow for a named environment.

What is trust-aware noise filtering in error triage?▼

Trust-aware noise filtering applies low-trust filters to raw watchdog errors before promoting tickets to lane-ready status. It uses trust thresholds, velocity boosts, and cross-environment escalation signals to decide which errors become actionable tickets.

How do error triage cycles send Slack notifications for promoted tickets?▼

After fingerprinting, deduplication, and promotion rules apply, the triage cycle sends notifications for promoted tickets. It also enforces Beads ticket label and lane transitions based on trust and thresholds, updating the checkpoint state.

Can I run error triage for multiple monitoring environments simultaneously?▼

The triage cycle runs one complete pass for a named environment at a time. It applies to monitoring and error-routing workflows for multiple configured environments, typically invoked by a watcher or manual run for debugging.

Why are duplicate error tickets still appearing after deduplication?▼

Duplicates may appear if structured record fields differ between entry points, altering the fingerprint hash. The deduplication requires matching stable hashes from identical structured fields to successfully group the same error.