Skip to content

docs: Add troubleshooting guide with 4-layer diagnostic model #31

@diberry

Description

@diberry

Problem

Users have no systematic way to debug Squad failures. The 4-layer model (Governance, Platform, LLM, State) is documented in issue comments but not in user-facing docs.

From

Flight + Procedures reviews of #21. Phase 2 in revised plan.

Proposed

Create docs/troubleshooting.md covering:

  • The 4-layer failure model with clear definitions
  • Investigation checklists per symptom (NOT blame tables)
  • squad doctor as first step for every problem
  • Common failure patterns with exact fix commands
  • 'My agent did nothing' / 'Decisions keep growing' / 'Agent gave wrong answer' walkthroughs

Key Constraint

Heuristics must be framed as 'investigation checklist' not 'root cause' — multiple layers can produce identical symptoms.

Owner

Procedures (Prompt Engineer) + PAO (DevRel)

Metadata

Metadata

Assignees

No one assigned

    Labels

    enhancementNew feature or requestgo:needs-researchNeeds investigationsquadSquad triage inbox — Lead will assign to a membersquad:fidoAssigned to FIDO (Quality Owner)

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions