CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

resolution-advisor

Provide Red Hat documentation-backed resolution recommendations for classified job errors. Use when: - After failure analysis and host fact inspection: "How do I fix this?" - "What does Red Hat recommend for this error?" - "What's the fix for privilege escalation timeout?" - "Is this a known AAP issue?" NOT for: analyzing events (use job-failure-analyzer first) or checking host facts (use host-fact-inspector first).

Quality

79%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./rh-automation/skills/resolution-advisor/SKILL.md
SKILL.md
Quality
Evals
Security

Resolution Advisor

Prerequisites

Required MCP Servers: None required (this skill provides advisory based on documentation)

When to Use This Skill

Use this skill when:

  • After job-failure-analyzer has classified the error type
  • After host-fact-inspector has correlated host facts (when applicable)
  • As the final step in the forensic-troubleshooter workflow
  • User asks for resolution recommendations

Do NOT use when:

  • Job events haven't been analyzed yet (run job-failure-analyzer first)
  • Host facts haven't been checked (run host-fact-inspector first if hosts are involved)
  • User wants to execute a job (use governance-executor skill)

Workflow

Step 1: Consult Error Classification and Troubleshooting Documentation

CRITICAL: Document consultation MUST happen BEFORE providing recommendations.

Document Consultation (REQUIRED - Execute FIRST):

  1. Action: Read error-classification.md using the Read tool to understand the error taxonomy, classification decision tree, and resolution path mapping
  2. Action: Read job-troubleshooting.md using the Read tool to understand failure pattern details and Red Hat source citations
  3. Output to user: "I consulted error-classification.md and job-troubleshooting.md to determine the resolution path based on Red Hat's troubleshooting guidance."

Step 2: Determine Resolution Path

Using the error classification from job-failure-analyzer and the correlation from host-fact-inspector, map to the resolution path from error-classification.md:

ClassificationResolution OwnerRed Hat Doc Reference
Platform - Host UnreachableNetwork/Infra teamAAP 2.6 Troubleshooting Guide
Platform - EE UnavailablePlatform AdminAAP 2.6 EE Guide
Platform - CapacityPlatform AdminAAP 2.5 Instance Groups (Ch. 17)
Code - Undefined VariablePlaybook DeveloperAnsible Variable Precedence docs
Code - Wrong PackagePlaybook DeveloperRHEL Package Management docs
Code - Syntax ErrorPlaybook DeveloperAnsible Playbook Guide
Config - Privilege EscalationOps TeamAAP 2.6 Troubleshooting Guide
Config - Credential MismatchOps TeamAAP 2.5 Security Best Practices
Config - Service FailureOps Teamsystemd documentation
Config - Missing CollectionPlatform AdminAAP 2.6 EE Guide

Step 3: Generate Resolution Advisory

Output format:

## Resolution Advisory

### Error Summary
**Classification**: [Platform / Code / Configuration] Error
**Pattern**: [Specific pattern from error-classification.md]
**Affected**: [hosts/tasks from failure analysis]

### Red Hat Guidance

Per Red Hat's *[Source Document]*:
> "[Direct quote or guidance from the cited Red Hat documentation]"

### Resolution Steps

**Resolution Owner**: [Network/Infra / Platform Admin / Playbook Developer / Ops Team]

1. [Specific step with Red Hat doc reference]
2. [Specific step with Red Hat doc reference]
3. [Verification step]

### Prevention

Per Red Hat's *[Source Document]* ([Section]):
> "[Guidance on preventing this type of error]"

**Recommended governance control**: [e.g., "Add check mode to catch this in dry runs", "Configure notifications for this failure type"]

### Related Governance Gaps

[If the error points to a governance gap from governance-readiness.md domains, note it]
- Example: "Privilege escalation failures may indicate missing credential separation (Domain 4: Credential Security)"
- Example: "Capacity exhaustion may indicate insufficient workload isolation (Domain 6: Workload Isolation)"

Step 4: Suggest Governance Improvements

If the failure correlates with a governance readiness domain, recommend the user run a governance assessment:

  • Platform errors (capacity, EE) → Domains 5 (EE) and 6 (Workload Isolation)
  • Configuration errors (credentials, privilege escalation) → Domains 3 (RBAC) and 4 (Credential Security)
  • Repeated failures without notification → Domain 2 (Notification Coverage)

Dependencies

Required MCP Servers

  • None (advisory-only skill based on documentation)

Required MCP Tools

  • None (advisory-only skill based on documentation)

Related Skills

  • job-failure-analyzer - Provides error classification input
  • host-fact-inspector - Provides host correlation input
  • governance-readiness-assessor - For governance gap follow-up
  • execution-summary - Audit trail

Reference Documentation

Example Usage

User: "How do I fix the package not found error from Job #4451?"

Agent:

  1. Reads error-classification.md and job-troubleshooting.md
  2. Reports: "I consulted error-classification.md and job-troubleshooting.md to determine the resolution path."
  3. Classification: Code Error - Wrong Package Name (Pattern 2b from error-classification.md)
  4. Provides resolution:
## Resolution Advisory

### Error Summary
**Classification**: Code Error
**Pattern**: Wrong Package Name (2b)
**Affected**: Host web-prod-01, Task "Install security package"

### Red Hat Guidance

Per Red Hat's *AAP 2.6 Troubleshooting Guide*:
> "Module failures typically indicate an issue with the playbook task definition."

### Resolution Steps

**Resolution Owner**: Playbook Developer

1. Verify the package name exists in the target OS repositories
2. Check `ansible_distribution` and `ansible_distribution_version` for the host
3. Update the playbook with the correct package name
4. Re-run with check mode to validate before full execution

### Prevention

Per Red Hat's *Controller Best Practices*:
> "Use separate inventories for production and development environments."

Test playbook changes in a development inventory first.
Repository

RHEcosystemAppEng/agentic-plugins

Path

rh-automation/skills/resolution-advisor/SKILL.md

Last updated
First committed

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.