Hk Arch Audit

/hk-arch-audit — Agent-Native Architecture Review

Conducts a comprehensive review of the codebase against agent-native architecture principles, launching parallel sub-agents for each principle and producing a scored report.

Context

$ARGUMENTS

Core Principles

| # | Principle | One-liner | |---|-----------|-----------| | 1 | Action Parity | Whatever the user can do, the agent can do | | 2 | Tools as Primitives | Tools provide capability, not behavior | | 3 | Context Injection | System prompt includes dynamic context about app state | | 4 | Shared Workspace | Agent and user work in the same data space | | 5 | CRUD Completeness | Every entity has full Create, Read, Update, Delete | | 6 | UI Integration | Agent actions immediately reflected in UI | | 7 | Capability Discovery | Users can discover what the agent can do | | 8 | Prompt-Native Features | Features are prompts defining outcomes, not code |

Workflow

Step 1: Load reference material

Read references/agent-prompts.md in this skill's directory to get the detailed prompt for each sub-agent. That file contains the exact instructions, search strategies, and output format for all 8 principle audits.

Step 2: Check for single-principle audit

If $ARGUMENTS specifies a single principle, only run that one sub-agent. Valid argument forms:

| Argument | Principle | |----------|-----------| | action parity or 1 | Action Parity | | tools or primitives or 2 | Tools as Primitives | | context or injection or 3 | Context Injection | | shared or workspace or 4 | Shared Workspace | | crud or 5 | CRUD Completeness | | ui or integration or 6 | UI Integration | | discovery or 7 | Capability Discovery | | prompt or features or 8 | Prompt-Native Features |

For single-principle audits, run one Explore sub-agent and present its findings directly — no summary table needed.

Step 3: Launch parallel sub-agents

Launch 8 parallel sub-agents using the Task tool with subagent_type: Explore, one per principle. Each agent should:

1. Enumerate all relevant instances in the codebase (user actions, tools, contexts, data stores, etc.)
2. Check compliance against the principle
3. Provide a specific score in X out of Y (percentage%) format
4. List specific gaps and recommendations

Read references/agent-prompts.md for the exact prompt to give each agent. Pass the full agent prompt from that file — don't summarize it.

Step 4: Compile summary report

After all agents complete, compile the summary using this template:

## Agent-Native Architecture Review: [Project Name]

### Overall Score Summary

| Core Principle | Score | Percentage | Status |
|----------------|-------|------------|--------|
| Action Parity | X/Y | Z% | STATUS |
| Tools as Primitives | X/Y | Z% | STATUS |
| Context Injection | X/Y | Z% | STATUS |
| Shared Workspace | X/Y | Z% | STATUS |
| CRUD Completeness | X/Y | Z% | STATUS |
| UI Integration | X/Y | Z% | STATUS |
| Capability Discovery | X/Y | Z% | STATUS |
| Prompt-Native Features | X/Y | Z% | STATUS |

**Overall Agent-Native Score: X%**

### Status Legend
- Excellent (80%+)
- Partial (50-79%)
- Needs Work (<50%)

### Top 10 Recommendations by Impact

| Priority | Action | Principle | Effort |
|----------|--------|-----------|--------|

### What's Working Well

[List top 5 strengths]

Success criteria

• All 8 sub-agents complete their audits (or the single requested one)
• Each principle has a specific numeric score (X/Y format)
• Summary table shows all scores and status indicators
• Top 10 recommendations are prioritized by impact
• Report identifies both strengths and gaps

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: deepklarity
• Source: deepklarity/harness-kit
• License: MIT

Install and usage instructions live in the source repository linked above.