Agent Skills: Audit Layer

Audit a codebase for agentic layer coverage and identify investment opportunities. Use to assess agentic maturity and find gaps.

UncategorizedID: melodic-software/claude-code-plugins/audit-layer

Install this agent skill to your local

pnpm dlx add-skill https://github.com/melodic-software/claude-code-plugins/tree/HEAD/plugins/tac/skills/audit-layer

Skill Files

Browse the full folder contents for audit-layer.

Download Skill

Loading file tree…

plugins/tac/skills/audit-layer/SKILL.md

Skill Metadata

Name
audit-layer
Description
Audit a codebase for agentic layer coverage and identify investment opportunities. Use to assess agentic maturity and find gaps.

Audit Layer

Audit a codebase for agentic layer coverage and identify investment opportunities.

Arguments

  • $ARGUMENTS: Target directory to audit (defaults to current directory)

Instructions

You are auditing a codebase to assess its agentic layer maturity.

Step 1: Check Commands Directory

Look for .claude/commands/ or equivalent:

# Check existence
ls -la .claude/commands/ 2>/dev/null || echo "Not found"

# List templates if exists
ls .claude/commands/*.md 2>/dev/null | wc -l

Score: 20 points if exists with 3+ templates

Step 2: Check Specs Directory

Look for specs/ or equivalent:

# Check existence and count
ls specs/*.md 2>/dev/null | wc -l

Score: 15 points if exists with specs

Step 3: Check ADW Directory

Look for adws/ or equivalent:

# Check for core module
ls adws/adw_modules/agent.py 2>/dev/null

# Count workflow scripts
ls adws/adw_*.py 2>/dev/null | wc -l

Score: 25 points if adws/ exists, +20 if agent.py exists

Step 4: Check Hooks

Look for .claude/hooks/:

ls .claude/hooks/*.py 2>/dev/null | wc -l

Score: 10 points if hooks exist

Step 5: Check Observability

Look for agents/ output directory:

ls -d agents/*/ 2>/dev/null | wc -l

Score: 5 points if agent output exists

Step 6: Check Worktrees

Look for trees/ isolation:

ls -d trees/*/ 2>/dev/null | wc -l
git worktree list 2>/dev/null | wc -l

Score: 5 points if worktrees exist

Scoring

| Component | Max Points | | --- | --- | | .claude/commands/ | 20 | | specs/ | 15 | | adws/ | 25 | | adw_modules/agent.py | 20 | | hooks/ | 10 | | agents/ | 5 | | trees/ | 5 | | Total | 100 |

Maturity Levels

| Score | Level | Recommendation | | --- | --- | --- | | 0-20 | None | Start with minimum viable layer | | 21-40 | Basic | Add composed workflows | | 41-60 | Developing | Add hooks and triggers | | 61-80 | Advanced | Add worktree isolation | | 81-100 | Complete | Focus on optimization |

Output

Provide audit report:

## Agentic Layer Audit Report

**Directory:** {target}
**Date:** {today}
**Score:** {score}/100
**Level:** {level}

### Components Found
- [x/o] .claude/commands/ ({count} templates)
- [x/o] specs/ ({count} specs)
- [x/o] adws/ ({count} workflows)
- [x/o] adw_modules/agent.py
- [x/o] hooks/ ({count} hooks)
- [x/o] agents/ (output directory)
- [x/o] trees/ (worktree isolation)

### Gaps Identified
1. {missing component}
2. {missing component}

### Recommended Investments
1. {next investment}
2. {next investment}

### Time Investment Analysis
- Estimated current: {percent}% on agentic layer
- Target: 50%+ on agentic layer

Notes

  • Higher scores indicate more mature agentic layers
  • Focus investments on highest-impact gaps first
  • Target is 50%+ engineering time on agentic layer