convez.ai
MCP Server
itlightning

sparklogs-ai-plugins

itlightning/sparklogs-ai-plugins

Ver no GitHub
BomPluginScore · de 0 a 100

SparkLogs helps managed service providers and IT pros investigate client issues with evidence instead of guesswork. This repository packages the SparkLogs AI skills, commands, subagents, and MCP configuration for Claude Code, Cursor, and Codex.

Instalação

$npx convez-ai itlightning/sparklogs-ai-plugins

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
0
Forks
0
Acompanhando
0
Saúde
75%
Último commit
há 5 dias
Versão
v1.12.12
Issues abertas
0
Contribuidores
1

README

A tradução para o português deste README ainda está na fila.

SparkLogs AI Plugins

SparkLogs helps managed service providers investigate client issues with evidence instead of guesswork. This repository packages the SparkLogs AI skills, commands, subagents, and MCP configuration for Claude Code and Cowork, Cursor, and Codex.

The plugin gives your AI assistant a SparkLogs investigation workflow:

  • /sparklogs-investigate gathers evidence from SparkLogs telemetry and produces a factual system condition summary with citations.
  • /sparklogs-analyze-cause is an explicit second step that turns prior findings into candidate hypotheses with confirm/refute steps.
  • /sparklogs-summary re-renders an existing investigation summary for ticket updates or customer communication.
  • /sparklogs-explain explains the evidence behind a specific claim or finding.

The plugin is read-only. It queries SparkLogs through your configured SparkLogs MCP endpoint and does not remediate, change customer systems, or make final operational decisions.

Why MSPs Should Care

Expected outcomes:

  • Faster yet much deeper first-pass summaries for tickets and escalations.
  • More consistent investigation steps across technicians.
  • Clear visibility into what was checked and what remains outside SparkLogs visibility.
  • Better handoff notes for senior engineers and customer-facing updates.
  • Cause analysis that does not overreach and avoids unchecked conclusions.

Supported Hosts

Claude Code (and Cowork) and Cursor are the recommended Foundry hosts. Codex is supported via repo/local marketplace installation while public directory flows mature. A neutral Agent Skills package is generated under plugins/generic/sparklogs/ on dist for hosts that can import skills from a folder.

Install And Update

This repository uses a two-branch publishing model: dist is the generated default branch for marketplace-ready plugin packages, and source is the authoring branch for contributions.

Install from the GitHub repository marketplace root, not from a raw marketplace.json URL.

Trust And Safety

SparkLogs AI assists investigation; the MSP technician remains accountable for operational decisions. Every factual claim should cite evidence. Cause analysis is opt-in and labeled as speculative. Read the AI trust principles.

Contributing

The default branch dist is generated. Please open pull requests against source, not dist. See CONTRIBUTING.md and AGENTS.md.

License

This repository is licensed under the Apache 2.0 license (see LICENSE).

Trademarks

"SparkLogs" and "IT Lightning" are trademarks of IT Lightning, LLC.

Ver o original no GitHub

Bom para

Relacionados