levnikolaevich/ln-627-diagnosability-auditor

Paths: File paths (references/, ../ln-*) are relative to this skill directory.

Diagnosability Auditor (L3 Worker)

Type: L3 Worker

Specialized worker auditing whether operators can diagnose incidents.

Purpose & Scope

• Audit diagnosability (Category 10: Medium Priority)
• Check structured logs, metrics, tracing, correlation IDs, and log levels
• Emit ADD_DIAGNOSTIC_SIGNAL, STRUCTURE_LOGS, or PROPAGATE_CORRELATION
• Calculate compliance score (X/10)

Inputs

MANDATORY READ: Load references/audit_worker_core_contract.md and references/mcp_tool_preferences.md. Tool policy: You may run as an isolated subagent where host AGENTS.md is not in scope, so default to hex-line MCP first for file reads, search, and edits. Load references/mcp_integration_patterns.md only when MCP behavior is unclear.

Receives contextStore with tech stack, framework, codebase root, output_dir.

Use hex-graph first when traces, call paths, or cross-file references materially improve the audit. Use hex-line first for local code reads when available. If MCP is unavailable, unsupported, or not indexed, continue with built-in Read/Grep/Glob/Bash and state the fallback in the report.

Workflow

Detection policy: use two-layer detection (candidate scan, then context verification); load references/two_layer_detection.md only when the verification method is ambiguous.

1. Parse context + output_dir
2. Determine project type (Layer 2 pre-check): Is this a web service (all checks apply), CLI tool (health/probes not applicable), or library (most checks optional)? Adjust applicable checks accordingly.
3. Check observability patterns (Layer 1: grep)
4. Analyze context per candidate (Layer 2):
   • Structured logging: is this a library (no logging OK) or a service (logging required)?
   • Request tracing: monolith -> less needed. Microservice -> critical
5. Collect confirmed findings
6. Calculate score
7. Write Report: Build full markdown report in memory per references/templates/audit_worker_report_template.md, write to {output_dir}/ln-627--global.md in single Write call
8. Return Summary: Return minimal summary

Audit Rules

1. Structured Logging

Detection:

• Grep for console.log (unstructured)
• Check for proper logger: winston, pino, logrus, zap

Severity:

• MEDIUM: Production code using console.log
• LOW: Dev code using console.log

Recommendation: Use structured logger (winston, pino)

Effort: M (add logger, replace calls)

2. Correlation IDs

Detection:

• Check for request ID/correlation ID middleware
• Verify IDs appear in logs and are propagated to outbound calls

Severity:

• MEDIUM: No correlation IDs in request-handling services

Recommendation: Add request ID middleware and include correlation ID in structured logs

Effort: M

3. Metrics Collection

Detection:

• Check for Prometheus client, StatsD, CloudWatch
• Grep for metric recording: histogram, counter

Severity:

• MEDIUM: No metrics instrumentation

Recommendation: Add Prometheus metrics

Effort: M (instrument code)

4. Request Tracing

Detection:

• Check for correlation IDs in logs
• Verify trace propagation (OpenTelemetry, Zipkin)

Severity:

• MEDIUM: No correlation IDs (hard to debug distributed systems)

Recommendation: Add request ID middleware

Effort: M (add middleware, propagate IDs)

5. Log Levels

Detection:

• Check if logger supports levels (info, warn, error, debug)
• Verify proper level usage

Severity:

• LOW: Only error logging (insufficient visibility)

Recommendation: Add info/debug logs

Effort: S (add log statements)

Scoring Algorithm

MANDATORY READ: Load references/audit_scoring.md.

Output Format

MANDATORY READ: Load references/templates/audit_worker_report_template.md.

Write JSON summary per references/audit_summary_contract.md. In managed mode the caller passes both runId and summaryArtifactPath; in standalone mode the worker generates its own run-scoped artifact path per shared contract.

Write report to {output_dir}/ln-627--global.md with category: "Diagnosability" and checks: structured_logging, correlation_ids, metrics_collection, request_tracing, log_levels.

Return summary per references/audit_summary_contract.md.

When summaryArtifactPath is absent, write the standalone runtime summary under .hex-skills/runtime-artifacts/runs/{run_id}/evaluation-worker/{worker}--{identifier}.json and optionally echo the same summary in structured output.

Report written: .hex-skills/runtime-artifacts/runs/{run_id}/audit-report/ln-627--global.md
Score: X.X/10 | Issues: N (C:N H:N M:N L:N)

Reference Files

• Audit output schema: references/audit_output_schema.md

Critical Rules

Apply the already-loaded references/audit_worker_core_contract.md.

• Do not auto-fix: Report only, never inject logging or endpoints
• Framework-aware detection: Adapt patterns to project's tech stack (winston/pino for Node, logrus/zap for Go, etc.)
• Effort realism: S = <1h, M = 1-4h, L = >4h
• Exclusions: Skip test files for console.log detection, skip dev-only scripts
• Context-sensitive severity: console.log in production code = MEDIUM, in dev utilities = LOW
• Unique angle: Audit diagnostic signal only. Liveness/readiness probes and shutdown behavior belong to lifecycle/config audit.
• Action required: Every finding uses ADD_DIAGNOSTIC_SIGNAL, STRUCTURE_LOGS, or PROPAGATE_CORRELATION.

Definition of Done

Apply the already-loaded references/audit_worker_core_contract.md.

• [ ] contextStore parsed (tech stack, framework, output_dir)
• [ ] All 5 checks completed (structured logging, correlation IDs, metrics, request tracing, log levels)
• [ ] Findings collected with severity, location, effort, action, recommendation
• [ ] Score calculated per references/audit_scoring.md
• [ ] Report written to {output_dir}/ln-627--global.md (atomic single Write call)
• [ ] Summary written per contract

---

Version: 3.0.0 Last Updated: 2025-12-23