Agent Skills: Inline Code Documentation Auditor (L3 Worker)

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

Inline Code Documentation Auditor (L3 Worker)

Type: L3 Worker

Specialized worker auditing inline code documentation quality: comments, docstrings, and language-specific documentation blocks.

Purpose & Scope

• Audit inline code documentation for quality and compliance across 6 categories
• Universal for any tech stack (auto-detect comment syntax)
• Return structured findings to coordinator with severity, location, recommendations
• Calculate compliance score (X/10) for Inline Code Documentation category
• Scope is limited to comments/docstrings/JSDoc/XML docs
• Out of scope: code design quality, naming quality, test quality, architecture quality, or feature correctness except where comments contradict code

Inputs

MANDATORY READ: Load references/audit_worker_core_contract.md and references/mcp_tool_preferences.md.

Receives contextStore with: tech_stack, project_root, output_dir.

Workflow

1. Parse Context: Extract tech stack, project root, output_dir from contextStore
2. Scan: Find all source files (use tech_stack for detection) Hex-line primary path: Use outline(file_path) and discovery-first read_file() for code files before analyzing comments. Use edit_ready=true, verbosity="full" only if the audit turns into a follow-up edit. Do not use hex-graph here - comment quality is a code-reading problem, not a semantic graph problem.
3. Extract: Parse inline comments + docstrings/JSDoc
4. Audit: Run 6 category checks (see Audit Categories below)
5. Collect Findings: Record each violation with severity, location (file:line), effort estimate (S/M/L), recommendation
6. Calculate Score: Count violations by severity, calculate compliance score (X/10)
7. Write Report: Build full markdown report per references/templates/audit_worker_report_template.md, write to {output_dir}/ln-613--global.md in single Write call
8. Return Summary: Return minimal summary to coordinator (see Output Format)

Audit Categories

| # | Category | What to Check | |---|----------|---------------| | 1 | WHY not WHAT | Comments explain rationale, not obvious code behavior; no restating code | | 2 | Density (15-20%) | Comment-to-code ratio within range; not over/under-commented | | 3 | No Forbidden Content | No dates/authors; no historical notes; no code examples in comments | | 4 | Docstrings Quality | Match function signatures; parameters documented; return types accurate | | 5 | Actuality | Comments match code behavior; no stale references; examples runnable | | 6 | Legacy Cleanup | No TODO without context; no commented-out code; no unsupported notes |

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-613--global.md with category: "Inline Code Documentation" and checks: why_not_what, density, forbidden_content, docstrings_quality, actuality, legacy_cleanup.

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-613--global.md
Score: X.X/10 | Issues: N (C:N H:N M:N L:N)

Severity mapping:

| Issue Type | Severity | |------------|----------| | Author names, dates in comments | CRITICAL | | Commented-out code blocks | HIGH | | Stale/outdated comments | HIGH | | Obvious WHAT comments | MEDIUM | | Density deviation >5% | MEDIUM | | Minor density deviation | LOW |

Critical Rules

Apply the already-loaded references/audit_worker_core_contract.md.

• Do not auto-fix: Report violations only; coordinator aggregates for user
• Fix code, not rules: NEVER modify rules files (*_rules.md, *_standards.md) to make violations pass
• Code is truth: When comment contradicts code, flag the documentation for update; do not turn this into a general code-quality review
• WHY > WHAT: Comments explaining obvious behavior should be removed
• Universal: Works with any language; detect comment syntax automatically
• Location precision: Always include file:line for programmatic navigation

Definition of Done

Apply the already-loaded references/audit_worker_core_contract.md.

• [ ] contextStore parsed successfully (including output_dir)
• [ ] All source files scanned (tech stack from contextStore)
• [ ] All 6 categories audited
• [ ] Findings collected with severity, location, effort, recommendation
• [ ] Score calculated using penalty algorithm
• [ ] Report written to {output_dir}/ln-613--global.md (atomic single Write call)
• [ ] Summary written per contract

Reference Files

• Comment rules and patterns: references/comments_rules.md

Version: 4.0.0 Last Updated: 2026-03-01