@dzhechkov/harness-cli 0.2.0 → 0.2.2
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +226 -16
- package/package.json +9 -3
package/README.md
CHANGED
|
@@ -1,27 +1,237 @@
|
|
|
1
1
|
# @dzhechkov/harness-cli
|
|
2
2
|
|
|
3
|
-
The **`dz`**
|
|
4
|
-
argv shell over [`@dzhechkov/harness-core`](../harness-core).
|
|
3
|
+
The **`dz`** CLI — the main entry point to the DZ Harness Hub. Install AI skills for **Claude Code, Codex, OpenCode, Hermes** from a single command.
|
|
5
4
|
|
|
6
|
-
##
|
|
5
|
+
## Install
|
|
7
6
|
|
|
7
|
+
```bash
|
|
8
|
+
npm install -g @dzhechkov/harness-cli
|
|
8
9
|
```
|
|
9
|
-
|
|
10
|
-
|
|
11
|
-
|
|
12
|
-
|
|
13
|
-
|
|
10
|
+
|
|
11
|
+
## Quick Start
|
|
12
|
+
|
|
13
|
+
```bash
|
|
14
|
+
# Install a preset (curated skill set) for your platform:
|
|
15
|
+
dz init --target claude-code --preset meta
|
|
16
|
+
|
|
17
|
+
# That's it — Claude Code will auto-discover the installed skills.
|
|
18
|
+
```
|
|
19
|
+
|
|
20
|
+
## Presets vs Individual Skills
|
|
21
|
+
|
|
22
|
+
| Approach | When to use | Example |
|
|
23
|
+
|----------|------------|---------|
|
|
24
|
+
| **Preset** | Want a curated set of skills that work together | `dz init --target claude-code --preset keysarium` |
|
|
25
|
+
| **--select** | Want specific skills by name | `dz init --target claude-code --select explore,feature-adr` |
|
|
26
|
+
| **Standalone npx** | Want a full toolkit with its own CLI | `npx @dzhechkov/keysarium init` |
|
|
27
|
+
|
|
28
|
+
### Available Presets (7)
|
|
29
|
+
|
|
30
|
+
| Preset | Skills | Description |
|
|
31
|
+
|--------|--------|-------------|
|
|
32
|
+
| `meta` | 3 | Development process (explore, feature-adr, knowledge-extractor) |
|
|
33
|
+
| `qe-engineer` | 8 | Quality engineering (test-gen, coverage, chaos, defect, ...) |
|
|
34
|
+
| `bto` | 1 | Build-Benchmark-Test-Optimize pipeline |
|
|
35
|
+
| `health` | 8 | Medical AI (diagnostics, drugs, labs, clinical decisions) |
|
|
36
|
+
| `keysarium` | 9 | Full research toolkit (feature-adr, presentation, reverse-eng) |
|
|
37
|
+
| `p-replicator` | 10 | AI product development (/replicate, SPARC PRD, pipeline-forge) |
|
|
38
|
+
| `feature-adr` | 5 | Feature pipeline (feature-adr, explore, frontend-design) |
|
|
39
|
+
|
|
40
|
+
### Standalone Packages (install via npx, no dz CLI needed)
|
|
41
|
+
|
|
42
|
+
```bash
|
|
43
|
+
npx @dzhechkov/keysarium init # full research toolkit
|
|
44
|
+
npx @dzhechkov/p-replicator init # AI product development
|
|
45
|
+
npx @dzhechkov/health-advisor init # medical AI (25 skills)
|
|
46
|
+
npx @dzhechkov/skills-bto init # BTO benchmarking
|
|
47
|
+
npx @dzhechkov/skills-feature-adr init # 11-step feature pipeline
|
|
48
|
+
npx @dzhechkov/skills-edu-site init # gamified edu site generator
|
|
49
|
+
npx @dzhechkov/skills-transcript-site init # transcript → interactive site
|
|
50
|
+
npx @dzhechkov/skills-analyst-manual init # 3-phase analyst composite
|
|
51
|
+
```
|
|
52
|
+
|
|
53
|
+
**Difference:** `dz init --preset` installs individual skills from `.claude/skills/` source into a target platform tree. Standalone `npx` packages have their own CLI and install a complete toolkit with commands, rules, shards, and agents — a richer but self-contained experience.
|
|
54
|
+
|
|
55
|
+
## All Commands (11)
|
|
56
|
+
|
|
57
|
+
```
|
|
58
|
+
dz init --target <name> [--preset <name>] [--select id,id,...] [--force]
|
|
59
|
+
dz verify [--skills-dir <dir>] [--target <name>]
|
|
60
|
+
dz sync [--canonical <dir>] [--project <dir>] [--dry-run] [--force]
|
|
61
|
+
dz update (alias for sync)
|
|
62
|
+
dz list [--skills-dir <dir>]
|
|
63
|
+
dz info --id <skill-id> [--skills-dir <dir>]
|
|
64
|
+
dz migrate [--project <dir>]
|
|
65
|
+
dz workflow --task <name> [--dry-run]
|
|
66
|
+
dz doctor [--project <dir>]
|
|
67
|
+
dz roam [--apply] [--slug <slug>]
|
|
14
68
|
dz help
|
|
15
69
|
```
|
|
16
70
|
|
|
17
|
-
|
|
18
|
-
|
|
19
|
-
|
|
20
|
-
|
|
21
|
-
|
|
71
|
+
### Targets (4 platforms)
|
|
72
|
+
|
|
73
|
+
| Target | Skills directory |
|
|
74
|
+
|--------|-----------------|
|
|
75
|
+
| `claude-code` | `.claude/skills/` |
|
|
76
|
+
| `codex` | `.agents/skills/` |
|
|
77
|
+
| `opencode` | `.opencode/skills/` |
|
|
78
|
+
| `hermes` | `.hermes/skills/` |
|
|
79
|
+
|
|
80
|
+
### Workflows (Opus 4.8+ dynamic workflows)
|
|
81
|
+
|
|
82
|
+
```bash
|
|
83
|
+
dz workflow --task coverage-lift # parallel coverage improvement
|
|
84
|
+
dz workflow --task mutation-kill # kill surviving mutants
|
|
85
|
+
dz workflow --task canonicalize # canonicalize new packages
|
|
86
|
+
dz workflow --task security-audit # adversarial security scan
|
|
87
|
+
```
|
|
88
|
+
|
|
89
|
+
## How it works
|
|
90
|
+
|
|
91
|
+
- `dz init` compiles canonical skills from the [agentskills.io](https://agentskills.io) standard into the target platform's layout
|
|
92
|
+
- Writing is **additive** — existing files are never overwritten without `--force`
|
|
93
|
+
- All 4 platform adapters produce **byte-identical** output (ADR-005)
|
|
94
|
+
- `dz doctor` runs 7 health checks (node version, adapters, config, SQLite, skills)
|
|
95
|
+
- `dz migrate` detects legacy keysarium/bto installations and recommends migration path
|
|
96
|
+
|
|
97
|
+
---
|
|
98
|
+
|
|
99
|
+
## Use Cases
|
|
100
|
+
|
|
101
|
+
### 1. Short-term product research (one-off study)
|
|
102
|
+
|
|
103
|
+
**Goal:** Quickly research a product idea, competitors, market — get a structured report.
|
|
104
|
+
|
|
105
|
+
```bash
|
|
106
|
+
# Option A: via dz CLI
|
|
107
|
+
dz init --target claude-code --preset meta
|
|
108
|
+
# Then in Claude Code:
|
|
109
|
+
# /explore "Research the market for AI-powered code review tools"
|
|
110
|
+
# /feature-adr "Summarize findings into an ADR"
|
|
111
|
+
|
|
112
|
+
# Option B: via keysarium (full 7-phase pipeline)
|
|
113
|
+
npx @dzhechkov/keysarium init
|
|
114
|
+
# Then in Claude Code:
|
|
115
|
+
# /casarium "AI-powered code review tools — market analysis"
|
|
116
|
+
# → Phase 0: Discovery → Phase 1: Exploration → Phase 2: Paranoid Research
|
|
117
|
+
# → Phase 3: Solution Design → Phase 4: Architecture → Phase 5: Presentation
|
|
118
|
+
```
|
|
119
|
+
|
|
120
|
+
**What you get:**
|
|
121
|
+
- `meta` preset: `/explore` clarifies the problem → `/feature-adr` structures findings as ADR decisions
|
|
122
|
+
- `keysarium`: full 7-phase pipeline with dream cycles, background workers, and presentation generation
|
|
123
|
+
|
|
124
|
+
**Best for:** Quick study (hours), competitive analysis, technology evaluation.
|
|
125
|
+
|
|
126
|
+
---
|
|
127
|
+
|
|
128
|
+
### 2. Long-term product research (evolving over time)
|
|
129
|
+
|
|
130
|
+
**Goal:** Continuously gather data, add new sources, and "recalculate" the product vision as insights accumulate.
|
|
131
|
+
|
|
132
|
+
```bash
|
|
133
|
+
# Install keysarium + memory layer
|
|
134
|
+
npx @dzhechkov/keysarium init
|
|
135
|
+
npm install -g @dzhechkov/harness-cli
|
|
136
|
+
dz init --target claude-code --preset meta
|
|
137
|
+
|
|
138
|
+
# Additionally install the memory package for persistent learning:
|
|
139
|
+
# (already included if using keysarium)
|
|
140
|
+
```
|
|
141
|
+
|
|
142
|
+
**Workflow — iterative research cycles:**
|
|
143
|
+
|
|
144
|
+
```
|
|
145
|
+
Week 1: /casarium "Product X — initial research"
|
|
146
|
+
→ researches/ directory created with findings
|
|
147
|
+
→ .keysarium/memory/ stores patterns + reward scores
|
|
148
|
+
|
|
149
|
+
Week 2: Add new data → /casarium "Product X — update with Q2 metrics"
|
|
150
|
+
→ Memory recalls Week 1 patterns (reward-calibrated learning)
|
|
151
|
+
→ New findings merged with existing, conflicts resolved
|
|
152
|
+
→ Product vision "recalculated" based on accumulated evidence
|
|
153
|
+
|
|
154
|
+
Week N: /casarium "Product X — pivot analysis after customer feedback"
|
|
155
|
+
→ Full history available via memory layer
|
|
156
|
+
→ /harvest extracts reusable knowledge patterns
|
|
157
|
+
```
|
|
158
|
+
|
|
159
|
+
**Key features for long-term research:**
|
|
160
|
+
- **Reward-calibrated memory** (`@dzhechkov/memory` Reflexion): each checkpoint response trains the system — "ок" = excellent (1.0), feedback = good (0.7), rework = needs_work (0.3)
|
|
161
|
+
- **Agent SDK Dreaming**: between sessions, patterns are consolidated and distilled
|
|
162
|
+
- **`/harvest`** (knowledge-extractor skill): extracts reusable patterns from completed research into `lib/` templates
|
|
163
|
+
- **SQLite + FTS5 backend**: scales to 100k+ records with full-text search across all research sessions
|
|
164
|
+
|
|
165
|
+
**Best for:** Product strategy over months, continuous market monitoring, evolving product vision.
|
|
166
|
+
|
|
167
|
+
---
|
|
168
|
+
|
|
169
|
+
### 3. Product research + working prototype
|
|
170
|
+
|
|
171
|
+
**Goal:** Research the product AND build a functional prototype.
|
|
172
|
+
|
|
173
|
+
#### Option A: Sequential — research first, then code
|
|
174
|
+
|
|
175
|
+
```bash
|
|
176
|
+
# Step 1: Install research + development presets
|
|
177
|
+
npx @dzhechkov/keysarium init
|
|
178
|
+
# OR:
|
|
179
|
+
dz init --target claude-code --preset keysarium
|
|
180
|
+
|
|
181
|
+
# Step 2: Research phase
|
|
182
|
+
# /casarium "SaaS platform for team retrospectives"
|
|
183
|
+
# → Phase 0-2: Discovery, Exploration, Paranoid Research
|
|
184
|
+
# → Phase 3: Solution Design (with CJM prototype)
|
|
185
|
+
# → Result: researches/<slug>/ with full analysis
|
|
186
|
+
|
|
187
|
+
# Step 3: Switch to development
|
|
188
|
+
dz init --target claude-code --preset feature-adr
|
|
189
|
+
|
|
190
|
+
# Step 4: Build using research outputs
|
|
191
|
+
# /feature-adr "Build the retrospective platform based on research in researches/<slug>/"
|
|
192
|
+
# → Step 0: Router classifies as L/XL
|
|
193
|
+
# → Step 1-5: Requirements, ADRs, DDD, Architecture (informed by research)
|
|
194
|
+
# → Step 6: Implementation plan
|
|
195
|
+
# → Step 7: Code generation (with /frontend-design for UI)
|
|
196
|
+
# → Step 8-9: QE review + fleet assessment
|
|
197
|
+
```
|
|
198
|
+
|
|
199
|
+
**What you get:** Research artifacts in `researches/`, then code in `features/<slug>/` + actual repository changes. Research directly feeds into ADR decisions.
|
|
200
|
+
|
|
201
|
+
#### Option B: Parallel — research and code simultaneously with p-replicator
|
|
202
|
+
|
|
203
|
+
```bash
|
|
204
|
+
# Install the full product development toolkit
|
|
205
|
+
npx @dzhechkov/p-replicator init
|
|
206
|
+
|
|
207
|
+
# Single pipeline: research → requirements → prototype
|
|
208
|
+
# /replicate "SaaS platform for team retrospectives"
|
|
209
|
+
# → Reverse-engineers similar products (reverse-engineering-unicorn)
|
|
210
|
+
# → Generates SPARC PRD (sparc-prd-mini)
|
|
211
|
+
# → Validates requirements (requirements-validator)
|
|
212
|
+
# → Creates the project structure (pipeline-forge)
|
|
213
|
+
# → Builds the prototype (cc-toolkit-generator-enhanced)
|
|
214
|
+
# → Reviews with brutal honesty (brutal-honesty-review)
|
|
215
|
+
```
|
|
216
|
+
|
|
217
|
+
**What you get:** A working prototype generated from research in a single `/replicate` pipeline run. Faster but less deep than Option A.
|
|
218
|
+
|
|
219
|
+
#### Comparison
|
|
220
|
+
|
|
221
|
+
| Aspect | Option A (Sequential) | Option B (p-replicator) |
|
|
222
|
+
|--------|----------------------|------------------------|
|
|
223
|
+
| **Research depth** | Deep (7-phase keysarium) | Moderate (reverse-engineering) |
|
|
224
|
+
| **Code quality** | High (11-step feature-adr + QE) | Good (pipeline-forge + review) |
|
|
225
|
+
| **Time** | Days to weeks | Hours to days |
|
|
226
|
+
| **Best for** | Complex products, regulated domains | MVPs, hackathons, quick validation |
|
|
227
|
+
| **Packages** | `keysarium` + `feature-adr` preset | `p-replicator` |
|
|
228
|
+
| **Research artifacts** | `researches/` directory | Embedded in PRD |
|
|
229
|
+
| **Code artifacts** | `features/<slug>/` + repo changes | Generated project |
|
|
230
|
+
|
|
231
|
+
**Tip:** For maximum rigor, combine both — use `p-replicator` for a quick prototype, then run `/feature-adr --full-qe-extended` on the generated code for production-grade quality engineering.
|
|
232
|
+
|
|
233
|
+
---
|
|
22
234
|
|
|
23
235
|
## Status
|
|
24
236
|
|
|
25
|
-
`
|
|
26
|
-
`--preset <name>` (Phase 10) selects a named skill set — see
|
|
27
|
-
`@dzhechkov/harness-presets`. The `paperclip` target is planned for a later phase.
|
|
237
|
+
`v0.2.1` — published on npm. Part of [DZ Harness Hub](https://github.com/djd1m/dz-harness-hub).
|
package/package.json
CHANGED
|
@@ -1,15 +1,21 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "@dzhechkov/harness-cli",
|
|
3
|
-
"version": "0.2.
|
|
4
|
-
"description": "The dz CLI
|
|
3
|
+
"version": "0.2.2",
|
|
4
|
+
"description": "The dz CLI — install AI skills for Claude Code, Codex, OpenCode, Hermes. 11 commands, 7 presets, 4 platform adapters.",
|
|
5
5
|
"type": "module",
|
|
6
6
|
"license": "MIT",
|
|
7
7
|
"author": "dzhechko",
|
|
8
8
|
"keywords": [
|
|
9
9
|
"agent-skills",
|
|
10
|
+
"agentskills.io",
|
|
10
11
|
"harness",
|
|
11
12
|
"cli",
|
|
12
|
-
"claude-code"
|
|
13
|
+
"claude-code",
|
|
14
|
+
"codex",
|
|
15
|
+
"opencode",
|
|
16
|
+
"hermes",
|
|
17
|
+
"ai-skills",
|
|
18
|
+
"cross-platform"
|
|
13
19
|
],
|
|
14
20
|
"main": "./dist/index.js",
|
|
15
21
|
"types": "./dist/index.d.ts",
|