@gonzih/skills-devops 1.0.0

package/README.md

@@ -0,0 +1,18 @@
+# @gonzih/skills-devops
+
+DevOps skills for Claude Code.
+
+## Install
+
+```bash
+npm install -g @gonzih/skills-devops
+```
+
+Skills are automatically copied to `~/.claude/skills/` on install.
+
+## Included Skills
+
+- **ci-cd-troubleshooter** — Diagnose and fix CI/CD pipeline failures
+- **runbook-writer** — Write operational runbooks for systems and services
+- **alert-triage** — Analyze and prioritize incoming alerts
+- **incident-commander** — Lead incident response and write postmortems

package/install.js

@@ -0,0 +1,13 @@
+import { copyFileSync, mkdirSync, readdirSync } from 'fs';
+import { join } from 'path';
+import { homedir } from 'os';
+
+const skillsDir = new URL('./skills/', import.meta.url).pathname;
+const targetDir = join(homedir(), '.claude', 'skills');
+
+for (const name of readdirSync(skillsDir)) {
+  const dest = join(targetDir, name);
+  mkdirSync(dest, { recursive: true });
+  copyFileSync(join(skillsDir, name, 'SKILL.md'), join(dest, 'SKILL.md'));
+  console.log(`Installed skill: ${name}`);
+}

package/package.json

@@ -0,0 +1 @@
+{"name":"@gonzih/skills-devops","version":"1.0.0","description":"DevOps skills for Claude Code","type":"module","scripts":{"postinstall":"node install.js"},"files":["skills/","install.js","README.md"],"keywords":["claude","mcp","skills","devops"],"license":"MIT"}

package/skills/alert-triage/SKILL.md

@@ -0,0 +1,25 @@
+---
+name: alert-triage
+description: Analyze and prioritize incoming alerts
+---
+
+## What
+Analyze a set of incoming monitoring alerts, determine severity and urgency, identify correlations, and produce a prioritized action plan.
+
+## How
+- Parse alert payloads (PagerDuty, Alertmanager, Datadog, CloudWatch, etc.)
+- Group related alerts by service, host, or time window to avoid alert storms
+- Assess business impact: user-facing vs. internal, revenue-critical vs. background
+- Identify the probable root alert vs. downstream noise
+- Recommend immediate actions vs. watch-and-wait
+
+## Workflow
+1. Paste or describe the active alerts
+2. Group and deduplicate correlated alerts
+3. Rank by impact: P1 (user-facing outage) → P4 (informational)
+4. Identify the root cause alert and downstream effects
+5. Output a triage summary: what to act on now, what to monitor, what to silence
+
+## Example
+User: "I have 47 alerts firing: disk full on db-01, high latency on api-gateway, 5xx spike on checkout"
+→ Identifies disk-full on db-01 as root cause driving the cascade, recommends clearing disk space first, provides ordered action plan.

package/skills/ci-cd-troubleshooter/SKILL.md

@@ -0,0 +1,25 @@
+---
+name: ci-cd-troubleshooter
+description: Diagnose and fix CI/CD pipeline failures
+---
+
+## What
+Diagnose failing CI/CD pipelines, identify root causes, and propose or apply fixes across GitHub Actions, GitLab CI, Jenkins, CircleCI, and similar systems.
+
+## How
+- Read pipeline config files and recent failure logs
+- Identify the failing step, error type, and likely cause
+- Check for common issues: missing secrets, dependency failures, flaky tests, resource limits, misconfigured caches
+- Propose targeted fixes with minimal blast radius
+- Verify the fix doesn't break other pipeline stages
+
+## Workflow
+1. Ask for the pipeline URL or paste the failure log
+2. Identify the failing job and step
+3. Classify the failure (config error, flaky test, infra issue, code bug)
+4. Propose a fix and explain the root cause
+5. Offer to apply the fix to the config file
+
+## Example
+User: "My GitHub Actions deploy job is failing with 'Error: Unable to locate executable file: docker'"
+→ Diagnose missing Docker setup step, suggest `docker/setup-buildx-action`, apply to workflow YAML.

package/skills/incident-commander/SKILL.md

@@ -0,0 +1,26 @@
+---
+name: incident-commander
+description: Lead incident response and write postmortems
+---
+
+## What
+Guide incident response from detection through resolution, coordinate communication, track timeline, and produce blameless postmortems.
+
+## How
+- Establish incident severity (SEV1–SEV4) and assign roles: IC, comms lead, tech lead
+- Maintain a live incident timeline with timestamped actions
+- Draft stakeholder updates at the right cadence and detail level
+- Track mitigation steps and confirm resolution criteria
+- After resolution: write a blameless postmortem with timeline, impact, root cause, and action items
+
+## Workflow
+1. **Declare**: Capture incident start time, description, and initial severity
+2. **Triage**: Identify affected systems, blast radius, and on-call owners
+3. **Coordinate**: Draft stakeholder update templates; track timeline entries
+4. **Mitigate**: Document each remediation step and its outcome
+5. **Resolve**: Confirm resolution criteria met; set follow-up review time
+6. **Postmortem**: Generate blameless postmortem doc with 5-whys root cause analysis and action items
+
+## Example
+User: "Production database is down, started 14:32 UTC, ~10k users affected"
+→ Declares SEV1, drafts initial stakeholder update, starts timeline, prompts for on-call contacts, and guides through mitigation steps to resolution and postmortem.

package/skills/runbook-writer/SKILL.md

@@ -0,0 +1,25 @@
+---
+name: runbook-writer
+description: Write operational runbooks for systems and services
+---
+
+## What
+Generate clear, actionable operational runbooks for services, infrastructure components, and recurring operational tasks.
+
+## How
+- Gather context: service name, tech stack, deployment platform, on-call audience
+- Structure runbooks with consistent sections: Overview, Prerequisites, Procedures, Rollback, Escalation
+- Write steps at the right level of detail for the target audience (SRE, developer, NOC)
+- Include verification commands after each major step
+- Add decision trees for common failure modes
+
+## Workflow
+1. Ask for the service/system name and what the runbook should cover
+2. Identify the target audience and their assumed knowledge level
+3. Draft the runbook with Overview → Prerequisites → Step-by-step Procedures → Verification → Rollback → Escalation
+4. Review for completeness: every step has a verification command, rollback is covered
+5. Save to the appropriate location (e.g., `docs/runbooks/<service>.md`)
+
+## Example
+User: "Write a runbook for restarting the payment-service in Kubernetes"
+→ Produces a runbook covering health checks, drain, rolling restart, verification, and rollback with `kubectl` commands.