nexus-dev-toolkit-jev 3.1.9__tar.gz

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.
Files changed (48) hide show
  1. nexus_dev_toolkit_jev-3.1.9/LICENSE +21 -0
  2. nexus_dev_toolkit_jev-3.1.9/PKG-INFO +262 -0
  3. nexus_dev_toolkit_jev-3.1.9/README.md +238 -0
  4. nexus_dev_toolkit_jev-3.1.9/nexus_cli.py +1031 -0
  5. nexus_dev_toolkit_jev-3.1.9/nexus_dev_toolkit_jev.egg-info/PKG-INFO +262 -0
  6. nexus_dev_toolkit_jev-3.1.9/nexus_dev_toolkit_jev.egg-info/SOURCES.txt +46 -0
  7. nexus_dev_toolkit_jev-3.1.9/nexus_dev_toolkit_jev.egg-info/dependency_links.txt +1 -0
  8. nexus_dev_toolkit_jev-3.1.9/nexus_dev_toolkit_jev.egg-info/entry_points.txt +3 -0
  9. nexus_dev_toolkit_jev-3.1.9/nexus_dev_toolkit_jev.egg-info/requires.txt +11 -0
  10. nexus_dev_toolkit_jev-3.1.9/nexus_dev_toolkit_jev.egg-info/top_level.txt +3 -0
  11. nexus_dev_toolkit_jev-3.1.9/nexus_server.py +16 -0
  12. nexus_dev_toolkit_jev-3.1.9/pyproject.toml +54 -0
  13. nexus_dev_toolkit_jev-3.1.9/setup.cfg +4 -0
  14. nexus_dev_toolkit_jev-3.1.9/tests/test_cli.py +645 -0
  15. nexus_dev_toolkit_jev-3.1.9/tests/test_iac_package_resolvers.py +220 -0
  16. nexus_dev_toolkit_jev-3.1.9/tests/test_server.py +21 -0
  17. nexus_dev_toolkit_jev-3.1.9/tests/test_typesafe_judgments.py +230 -0
  18. nexus_dev_toolkit_jev-3.1.9/tools/__init__.py +0 -0
  19. nexus_dev_toolkit_jev-3.1.9/tools/agents/__init__.py +0 -0
  20. nexus_dev_toolkit_jev-3.1.9/tools/agents/code-reviewer.md +86 -0
  21. nexus_dev_toolkit_jev-3.1.9/tools/agents/database-reviewer.md +95 -0
  22. nexus_dev_toolkit_jev-3.1.9/tools/agents/deployment-reviewer.md +97 -0
  23. nexus_dev_toolkit_jev-3.1.9/tools/agents/monitoring-reviewer.md +95 -0
  24. nexus_dev_toolkit_jev-3.1.9/tools/agents/performance-reviewer.md +98 -0
  25. nexus_dev_toolkit_jev-3.1.9/tools/agents/pr-verifier.md +63 -0
  26. nexus_dev_toolkit_jev-3.1.9/tools/epav/__init__.py +14 -0
  27. nexus_dev_toolkit_jev-3.1.9/tools/epav/arch_ingest.py +299 -0
  28. nexus_dev_toolkit_jev-3.1.9/tools/epav/graph_backend.py +88 -0
  29. nexus_dev_toolkit_jev-3.1.9/tools/epav/judgment.py +50 -0
  30. nexus_dev_toolkit_jev-3.1.9/tools/epav/package_resolver.py +624 -0
  31. nexus_dev_toolkit_jev-3.1.9/tools/epav/project_rules.py +267 -0
  32. nexus_dev_toolkit_jev-3.1.9/tools/epav/skills/__init__.py +0 -0
  33. nexus_dev_toolkit_jev-3.1.9/tools/epav/skills/apply.md +44 -0
  34. nexus_dev_toolkit_jev-3.1.9/tools/epav/skills/code-review.md +24 -0
  35. nexus_dev_toolkit_jev-3.1.9/tools/epav/skills/commit.md +48 -0
  36. nexus_dev_toolkit_jev-3.1.9/tools/epav/skills/create-pr.md +64 -0
  37. nexus_dev_toolkit_jev-3.1.9/tools/epav/skills/database-review.md +91 -0
  38. nexus_dev_toolkit_jev-3.1.9/tools/epav/skills/deployment-review.md +94 -0
  39. nexus_dev_toolkit_jev-3.1.9/tools/epav/skills/dev-tasks-planner.md +196 -0
  40. nexus_dev_toolkit_jev-3.1.9/tools/epav/skills/epav.md +46 -0
  41. nexus_dev_toolkit_jev-3.1.9/tools/epav/skills/evaluate.md +45 -0
  42. nexus_dev_toolkit_jev-3.1.9/tools/epav/skills/monitoring-review.md +91 -0
  43. nexus_dev_toolkit_jev-3.1.9/tools/epav/skills/performance-review.md +93 -0
  44. nexus_dev_toolkit_jev-3.1.9/tools/epav/skills/plan.md +46 -0
  45. nexus_dev_toolkit_jev-3.1.9/tools/epav/skills/pr-review.md +35 -0
  46. nexus_dev_toolkit_jev-3.1.9/tools/epav/skills/scaffold.md +419 -0
  47. nexus_dev_toolkit_jev-3.1.9/tools/epav/skills/validate.md +57 -0
  48. nexus_dev_toolkit_jev-3.1.9/tools/epav/task_loader.py +228 -0
@@ -0,0 +1,21 @@
1
+ MIT License
2
+
3
+ Copyright (c) 2026 Ronald dela Cruz
4
+
5
+ Permission is hereby granted, free of charge, to any person obtaining a copy
6
+ of this software and associated documentation files (the "Software"), to deal
7
+ in the Software without restriction, including without limitation the rights
8
+ to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
9
+ copies of the Software, and to permit persons to whom the Software is
10
+ furnished to do so, subject to the following conditions:
11
+
12
+ The above copyright notice and this permission notice shall be included in all
13
+ copies or substantial portions of the Software.
14
+
15
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
16
+ IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
17
+ FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
18
+ AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
19
+ LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
20
+ OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
21
+ SOFTWARE.
@@ -0,0 +1,262 @@
1
+ Metadata-Version: 2.4
2
+ Name: nexus-dev-toolkit-jev
3
+ Version: 3.1.9
4
+ Summary: Developer workflow toolkit for Claude Code — Day 0 scaffold + Day 1 EPAV (experimental TypeSafe/Jev build -- installs as nexus-jev/nexus-jev-mcp, side by side with the real nexus-dev-toolkit)
5
+ Author-email: Ronald dela Cruz <rcdelacruz@users.noreply.github.com>
6
+ License-Expression: MIT
7
+ Project-URL: Homepage, https://nexus.coderstudio.co
8
+ Project-URL: Documentation, https://nexus.coderstudio.co/docs
9
+ Project-URL: Repository, https://github.com/rcdelacruz/nexus-dev-toolkit
10
+ Project-URL: Issues, https://github.com/rcdelacruz/nexus-dev-toolkit/issues
11
+ Requires-Python: >=3.10
12
+ Description-Content-Type: text/markdown
13
+ License-File: LICENSE
14
+ Requires-Dist: mcp[cli]<2.0.0,>=1.27.1
15
+ Requires-Dist: typer>=0.12.0
16
+ Requires-Dist: rich>=13.0.0
17
+ Provides-Extra: dev
18
+ Requires-Dist: pytest>=8.0.0; extra == "dev"
19
+ Requires-Dist: pytest-asyncio>=0.23.0; extra == "dev"
20
+ Requires-Dist: typesafe-sdk<1.0.0,>=0.7.0; extra == "dev"
21
+ Provides-Extra: typesafe
22
+ Requires-Dist: typesafe-sdk<1.0.0,>=0.7.0; extra == "typesafe"
23
+ Dynamic: license-file
24
+
25
+ # nexus-dev-toolkit
26
+
27
+
28
+ Developer workflow toolkit for Claude Code and OpenCode. Gives any team a structured Day 0 scaffold and repeatable Day 1 feature cycle via the EPAV methodology.
29
+
30
+ **[Documentation](https://nexus.coderstudio.co/docs)**
31
+
32
+ ---
33
+
34
+ ## Why
35
+
36
+ Ad-hoc AI prompting doesn't scale. Every dev prompts differently, context drifts, and nobody knows what the AI was told last sprint.
37
+
38
+ nexus-dev-toolkit gives your team a single workflow:
39
+
40
+ - **Day 0** — scaffold the project once, production-grade, zero credentials needed
41
+ - **Day 1** — every feature follows the same four steps: evaluate → plan → apply → validate
42
+
43
+ Every skill, every rule, every pattern lives in the project repo — versioned, shared, and enforced.
44
+
45
+ ---
46
+
47
+ ## Setup
48
+
49
+ ### 1. Install prerequisites
50
+
51
+ **uv** (recommended package manager):
52
+ ```bash
53
+ curl -LsSf https://astral.sh/uv/install.sh | sh
54
+ ```
55
+
56
+ **Claude Code:**
57
+ ```bash
58
+ npm install -g @anthropic-ai/claude-code
59
+ claude login
60
+ ```
61
+
62
+ **graphify** _(strongly recommended)_ — builds a queryable knowledge graph of your codebase. EPAV skills query this graph to understand blast radius and cross-file dependencies before touching any code.
63
+
64
+ | | With graphify | Without graphify |
65
+ |---|---|---|
66
+ | Context loading | Scoped subgraph — fast, low token cost | Full file reads — slow, high token cost |
67
+ | Blast radius | Accurate, graph-backed | Best-effort, grep-based |
68
+ | Auto-update | PostToolUse hook keeps graph current | N/A |
69
+ | EPAV skills | Full capability | Degraded but functional |
70
+
71
+ ```bash
72
+ uv tool install graphifyy # PyPI package name has double y; command is "graphify"
73
+ graphify install # registers /graphify skill into Claude Code
74
+ ```
75
+
76
+ **codegraph** _(alternative to graphify)_ — a different knowledge-graph tool EPAV skills can use instead. Pick one, not both; `nexus doctor` warns if it finds graphs from both built in the same project.
77
+
78
+ ```bash
79
+ npm i -g @colbymchenry/codegraph
80
+ codegraph install # wires codegraph into Claude Code / OpenCode / etc.
81
+ ```
82
+
83
+ ### 2. Install nexus-dev-toolkit
84
+
85
+ ```bash
86
+ uv tool install nexus-dev-toolkit
87
+ ```
88
+
89
+ ### 3. Initialize your project
90
+
91
+ ```bash
92
+ cd my-project
93
+ nexus init . # Claude Code (default), prompts for graphify/codegraph/none + TypeSafe
94
+ nexus init . --tool opencode # OpenCode
95
+ nexus init . --graph-backend codegraph # skip the graph-backend prompt, use codegraph
96
+ nexus init . --typesafe # skip the TypeSafe prompt, enable it
97
+ ```
98
+
99
+ **TypeSafe/Jev** _(optional)_ — replaces a handful of hardcoded keyword-matching
100
+ heuristics in the EPAV tools (which architecture-doc section is about auth vs.
101
+ security, which of ~30 stack tags an arch doc mentions, which package manager
102
+ or IaC tool fits a free-text hint, which CSV column means "acceptance
103
+ criteria") with an actual judgment call. Every one of those tools works
104
+ exactly the same without it — TypeSafe is off by default, `pip install
105
+ nexus-dev-toolkit` never installs it, and every judgment falls back to the
106
+ original heuristic if it isn't enabled or the API call fails. Enable at init
107
+ with `--typesafe` (or answer `y` at the prompt), then set `TYPESAFE_API_KEY`
108
+ (get one at https://console.typesafe.ai/). `nexus doctor` reports whether
109
+ it's wired up and whether the key is set. See `tools/README.md` for how it's
110
+ implemented.
111
+
112
+ ### 4. Place your reference docs
113
+
114
+ Before running `/scaffold`, put everything Claude needs in `docs/`:
115
+
116
+ ```
117
+ docs/
118
+ ├── arch-docs/ ← architecture doc, ADRs
119
+ ├── figma/ ← Figma export ZIP
120
+ ├── brd/ ← Business Requirements Document
121
+ └── prd/ ← Product Requirements Document
122
+ ```
123
+
124
+ ### 5. Start Day 0
125
+
126
+ Open the project in Claude Code and run:
127
+ ```
128
+ /scaffold
129
+ ```
130
+
131
+ `/scaffold` runs EVALUATE first then stops. You drive each phase by typing the next command:
132
+
133
+ ```
134
+ /scaffold → (review) → /plan → (review) → /apply → (review) → /validate
135
+ ```
136
+
137
+ ### 6. Build the knowledge graph
138
+
139
+ After `/scaffold` completes, build whichever backend you chose at init:
140
+ ```
141
+ /graphify . # if you chose graphify
142
+ codegraph install && codegraph init # if you chose codegraph, run in your terminal
143
+ ```
144
+
145
+ This generates `graphify-out/graph.json` or `.codegraph/codegraph.db` — EPAV skills use whichever is present, or fall back to full file reads if neither is built yet. Run `nexus doctor` any time to check which backend is active.
146
+
147
+ ---
148
+
149
+ ## The Workflow
150
+
151
+ ### Day 0 — `/scaffold` (once per project)
152
+
153
+ Produces a production-grade project from your architecture document — plus a Figma design, for a project with a UI: correct stack, mock auth, mock data, design system, AGENTS.md — zero external dependencies. Runs `npm install && npm run dev` (or equivalent) from commit one. For an infrastructure-shaped project (Terraform/OpenTofu, Helm, Ansible), there's no Figma/UI/mock-auth step at all — it generates the IaC modules/playbooks/charts instead, validated with `plan`/`lint`/`template`, never `apply`d against real cloud accounts on Day 0.
154
+
155
+ ### Day 1 — EPAV (every feature, every sprint)
156
+
157
+ ```
158
+ /evaluate <task> → /plan → /apply → /validate
159
+ ```
160
+
161
+ Each step is a built-in skill in `.claude/commands/`. Every task starts from the dev tasks CSV. Every task ends with acceptance criteria verified.
162
+
163
+ ---
164
+
165
+ ## What `nexus init` Creates
166
+
167
+ `settings.json`/`graphify.js` below are only created if you pick graphify (the default) at the `--graph-backend` prompt. Pick codegraph instead and neither is scaffolded — codegraph wires and syncs itself via `codegraph install && codegraph init`.
168
+
169
+ **Claude Code** (`nexus init .`):
170
+ ```
171
+ .claude/
172
+ ├── commands/ ← EPAV skills + 5 reviewer skills (/code-review, etc.)
173
+ ├── agents/ ← 5 reviewer subagents (code-reviewer, database-reviewer, etc.)
174
+ └── settings.json ← PostToolUse: graphify auto-updates after every file edit (graphify backend only)
175
+ knowledge/
176
+ ├── rules/ ← coding standards, arch decisions
177
+ ├── patterns/ ← reusable implementation patterns
178
+ ├── prompts/dev/ ← task prompt templates
179
+ └── retros/ ← retrospective notes
180
+ .mcp.json ← MCP server config
181
+ ```
182
+
183
+ **OpenCode** (`nexus init . --tool opencode`):
184
+ ```
185
+ .opencode/
186
+ ├── commands/ ← same EPAV skills + reviewer skills
187
+ ├── agents/ ← same reviewer subagents (adapted for OpenCode)
188
+ └── plugins/
189
+ └── graphify.js ← tool.execute.after: graphify auto-update (graphify backend only)
190
+ knowledge/ ← same structure
191
+ opencode.json ← MCP server config
192
+ ```
193
+
194
+ ---
195
+
196
+ ## Commands
197
+
198
+ ```bash
199
+ nexus init . # Claude Code — set up .claude/ + knowledge/ + .mcp.json
200
+ nexus init . --tool opencode # OpenCode — set up .opencode/ + knowledge/ + opencode.json
201
+ nexus --version # show installed version
202
+ nexus update # update the CLI itself; asks whether to also sync in a real terminal
203
+ nexus update --sync # skip the prompt, always sync the current project's skills/agents too
204
+ nexus sync # sync built-in skills & agents to latest (custom files untouched)
205
+ nexus doctor # validate project setup
206
+
207
+ nexus skill list # list all skills in .claude/commands/
208
+ nexus skill add my-review # create a custom skill
209
+
210
+ nexus agent list # list all subagents in .claude/agents/
211
+ nexus agent add my-agent # create a custom subagent
212
+
213
+ nexus rule list # list all rules in knowledge/rules/
214
+ nexus rule add api-standards # create a project rule
215
+ ```
216
+
217
+ ---
218
+
219
+ ## MCP Server
220
+
221
+ `nexus init` writes `.mcp.json` automatically. For Claude Desktop, add manually:
222
+
223
+ ```json
224
+ {
225
+ "mcpServers": {
226
+ "nexus": {
227
+ "command": "uvx",
228
+ "args": ["--refresh", "--from", "nexus-dev-toolkit", "nexus-mcp"]
229
+ }
230
+ }
231
+ }
232
+ ```
233
+
234
+ Add `[typesafe]` to the package name (`"nexus-dev-toolkit[typesafe]"`) to enable
235
+ TypeSafe/Jev-powered detection — `nexus init --typesafe` does this for you.
236
+
237
+ ### MCP Tools
238
+
239
+ | Tool | Purpose |
240
+ |---|---|
241
+ | `ingest_architecture_doc` | Load arch doc → `knowledge/rules/arch-summary.md`; also classifies `project_shape` (web_app, mobile_app, backend_api, infrastructure, data_pipeline, cli_or_library) and `has_infrastructure_component`, so `/scaffold` knows whether to generate a UI shell, IaC deliverables, or both |
242
+ | `load_task` | Load a CSV task row into context — maps whatever column headers your CSV actually uses onto the standard task fields |
243
+ | `generate_project_rules` | Generate `AGENTS.md` from arch doc |
244
+ | `resolve_package_versions` | Resolve exact versions via the real tool — npm/pnpm/yarn/pub/go/cargo/maven/gradle/pip, or Terraform/OpenTofu, Helm, and Ansible for infrastructure-shaped projects |
245
+
246
+ ---
247
+
248
+ ## Custom Skills
249
+
250
+ ```bash
251
+ nexus skill add my-code-review
252
+ # Edit .claude/commands/my-code-review.md
253
+ # Type /my-code-review in Claude Code
254
+ ```
255
+
256
+ Custom skills live alongside built-in skills in `.claude/commands/` — versioned in your repo, shared across the team.
257
+
258
+ ---
259
+
260
+ ## License
261
+
262
+ [MIT](LICENSE)
@@ -0,0 +1,238 @@
1
+ # nexus-dev-toolkit
2
+
3
+
4
+ Developer workflow toolkit for Claude Code and OpenCode. Gives any team a structured Day 0 scaffold and repeatable Day 1 feature cycle via the EPAV methodology.
5
+
6
+ **[Documentation](https://nexus.coderstudio.co/docs)**
7
+
8
+ ---
9
+
10
+ ## Why
11
+
12
+ Ad-hoc AI prompting doesn't scale. Every dev prompts differently, context drifts, and nobody knows what the AI was told last sprint.
13
+
14
+ nexus-dev-toolkit gives your team a single workflow:
15
+
16
+ - **Day 0** — scaffold the project once, production-grade, zero credentials needed
17
+ - **Day 1** — every feature follows the same four steps: evaluate → plan → apply → validate
18
+
19
+ Every skill, every rule, every pattern lives in the project repo — versioned, shared, and enforced.
20
+
21
+ ---
22
+
23
+ ## Setup
24
+
25
+ ### 1. Install prerequisites
26
+
27
+ **uv** (recommended package manager):
28
+ ```bash
29
+ curl -LsSf https://astral.sh/uv/install.sh | sh
30
+ ```
31
+
32
+ **Claude Code:**
33
+ ```bash
34
+ npm install -g @anthropic-ai/claude-code
35
+ claude login
36
+ ```
37
+
38
+ **graphify** _(strongly recommended)_ — builds a queryable knowledge graph of your codebase. EPAV skills query this graph to understand blast radius and cross-file dependencies before touching any code.
39
+
40
+ | | With graphify | Without graphify |
41
+ |---|---|---|
42
+ | Context loading | Scoped subgraph — fast, low token cost | Full file reads — slow, high token cost |
43
+ | Blast radius | Accurate, graph-backed | Best-effort, grep-based |
44
+ | Auto-update | PostToolUse hook keeps graph current | N/A |
45
+ | EPAV skills | Full capability | Degraded but functional |
46
+
47
+ ```bash
48
+ uv tool install graphifyy # PyPI package name has double y; command is "graphify"
49
+ graphify install # registers /graphify skill into Claude Code
50
+ ```
51
+
52
+ **codegraph** _(alternative to graphify)_ — a different knowledge-graph tool EPAV skills can use instead. Pick one, not both; `nexus doctor` warns if it finds graphs from both built in the same project.
53
+
54
+ ```bash
55
+ npm i -g @colbymchenry/codegraph
56
+ codegraph install # wires codegraph into Claude Code / OpenCode / etc.
57
+ ```
58
+
59
+ ### 2. Install nexus-dev-toolkit
60
+
61
+ ```bash
62
+ uv tool install nexus-dev-toolkit
63
+ ```
64
+
65
+ ### 3. Initialize your project
66
+
67
+ ```bash
68
+ cd my-project
69
+ nexus init . # Claude Code (default), prompts for graphify/codegraph/none + TypeSafe
70
+ nexus init . --tool opencode # OpenCode
71
+ nexus init . --graph-backend codegraph # skip the graph-backend prompt, use codegraph
72
+ nexus init . --typesafe # skip the TypeSafe prompt, enable it
73
+ ```
74
+
75
+ **TypeSafe/Jev** _(optional)_ — replaces a handful of hardcoded keyword-matching
76
+ heuristics in the EPAV tools (which architecture-doc section is about auth vs.
77
+ security, which of ~30 stack tags an arch doc mentions, which package manager
78
+ or IaC tool fits a free-text hint, which CSV column means "acceptance
79
+ criteria") with an actual judgment call. Every one of those tools works
80
+ exactly the same without it — TypeSafe is off by default, `pip install
81
+ nexus-dev-toolkit` never installs it, and every judgment falls back to the
82
+ original heuristic if it isn't enabled or the API call fails. Enable at init
83
+ with `--typesafe` (or answer `y` at the prompt), then set `TYPESAFE_API_KEY`
84
+ (get one at https://console.typesafe.ai/). `nexus doctor` reports whether
85
+ it's wired up and whether the key is set. See `tools/README.md` for how it's
86
+ implemented.
87
+
88
+ ### 4. Place your reference docs
89
+
90
+ Before running `/scaffold`, put everything Claude needs in `docs/`:
91
+
92
+ ```
93
+ docs/
94
+ ├── arch-docs/ ← architecture doc, ADRs
95
+ ├── figma/ ← Figma export ZIP
96
+ ├── brd/ ← Business Requirements Document
97
+ └── prd/ ← Product Requirements Document
98
+ ```
99
+
100
+ ### 5. Start Day 0
101
+
102
+ Open the project in Claude Code and run:
103
+ ```
104
+ /scaffold
105
+ ```
106
+
107
+ `/scaffold` runs EVALUATE first then stops. You drive each phase by typing the next command:
108
+
109
+ ```
110
+ /scaffold → (review) → /plan → (review) → /apply → (review) → /validate
111
+ ```
112
+
113
+ ### 6. Build the knowledge graph
114
+
115
+ After `/scaffold` completes, build whichever backend you chose at init:
116
+ ```
117
+ /graphify . # if you chose graphify
118
+ codegraph install && codegraph init # if you chose codegraph, run in your terminal
119
+ ```
120
+
121
+ This generates `graphify-out/graph.json` or `.codegraph/codegraph.db` — EPAV skills use whichever is present, or fall back to full file reads if neither is built yet. Run `nexus doctor` any time to check which backend is active.
122
+
123
+ ---
124
+
125
+ ## The Workflow
126
+
127
+ ### Day 0 — `/scaffold` (once per project)
128
+
129
+ Produces a production-grade project from your architecture document — plus a Figma design, for a project with a UI: correct stack, mock auth, mock data, design system, AGENTS.md — zero external dependencies. Runs `npm install && npm run dev` (or equivalent) from commit one. For an infrastructure-shaped project (Terraform/OpenTofu, Helm, Ansible), there's no Figma/UI/mock-auth step at all — it generates the IaC modules/playbooks/charts instead, validated with `plan`/`lint`/`template`, never `apply`d against real cloud accounts on Day 0.
130
+
131
+ ### Day 1 — EPAV (every feature, every sprint)
132
+
133
+ ```
134
+ /evaluate <task> → /plan → /apply → /validate
135
+ ```
136
+
137
+ Each step is a built-in skill in `.claude/commands/`. Every task starts from the dev tasks CSV. Every task ends with acceptance criteria verified.
138
+
139
+ ---
140
+
141
+ ## What `nexus init` Creates
142
+
143
+ `settings.json`/`graphify.js` below are only created if you pick graphify (the default) at the `--graph-backend` prompt. Pick codegraph instead and neither is scaffolded — codegraph wires and syncs itself via `codegraph install && codegraph init`.
144
+
145
+ **Claude Code** (`nexus init .`):
146
+ ```
147
+ .claude/
148
+ ├── commands/ ← EPAV skills + 5 reviewer skills (/code-review, etc.)
149
+ ├── agents/ ← 5 reviewer subagents (code-reviewer, database-reviewer, etc.)
150
+ └── settings.json ← PostToolUse: graphify auto-updates after every file edit (graphify backend only)
151
+ knowledge/
152
+ ├── rules/ ← coding standards, arch decisions
153
+ ├── patterns/ ← reusable implementation patterns
154
+ ├── prompts/dev/ ← task prompt templates
155
+ └── retros/ ← retrospective notes
156
+ .mcp.json ← MCP server config
157
+ ```
158
+
159
+ **OpenCode** (`nexus init . --tool opencode`):
160
+ ```
161
+ .opencode/
162
+ ├── commands/ ← same EPAV skills + reviewer skills
163
+ ├── agents/ ← same reviewer subagents (adapted for OpenCode)
164
+ └── plugins/
165
+ └── graphify.js ← tool.execute.after: graphify auto-update (graphify backend only)
166
+ knowledge/ ← same structure
167
+ opencode.json ← MCP server config
168
+ ```
169
+
170
+ ---
171
+
172
+ ## Commands
173
+
174
+ ```bash
175
+ nexus init . # Claude Code — set up .claude/ + knowledge/ + .mcp.json
176
+ nexus init . --tool opencode # OpenCode — set up .opencode/ + knowledge/ + opencode.json
177
+ nexus --version # show installed version
178
+ nexus update # update the CLI itself; asks whether to also sync in a real terminal
179
+ nexus update --sync # skip the prompt, always sync the current project's skills/agents too
180
+ nexus sync # sync built-in skills & agents to latest (custom files untouched)
181
+ nexus doctor # validate project setup
182
+
183
+ nexus skill list # list all skills in .claude/commands/
184
+ nexus skill add my-review # create a custom skill
185
+
186
+ nexus agent list # list all subagents in .claude/agents/
187
+ nexus agent add my-agent # create a custom subagent
188
+
189
+ nexus rule list # list all rules in knowledge/rules/
190
+ nexus rule add api-standards # create a project rule
191
+ ```
192
+
193
+ ---
194
+
195
+ ## MCP Server
196
+
197
+ `nexus init` writes `.mcp.json` automatically. For Claude Desktop, add manually:
198
+
199
+ ```json
200
+ {
201
+ "mcpServers": {
202
+ "nexus": {
203
+ "command": "uvx",
204
+ "args": ["--refresh", "--from", "nexus-dev-toolkit", "nexus-mcp"]
205
+ }
206
+ }
207
+ }
208
+ ```
209
+
210
+ Add `[typesafe]` to the package name (`"nexus-dev-toolkit[typesafe]"`) to enable
211
+ TypeSafe/Jev-powered detection — `nexus init --typesafe` does this for you.
212
+
213
+ ### MCP Tools
214
+
215
+ | Tool | Purpose |
216
+ |---|---|
217
+ | `ingest_architecture_doc` | Load arch doc → `knowledge/rules/arch-summary.md`; also classifies `project_shape` (web_app, mobile_app, backend_api, infrastructure, data_pipeline, cli_or_library) and `has_infrastructure_component`, so `/scaffold` knows whether to generate a UI shell, IaC deliverables, or both |
218
+ | `load_task` | Load a CSV task row into context — maps whatever column headers your CSV actually uses onto the standard task fields |
219
+ | `generate_project_rules` | Generate `AGENTS.md` from arch doc |
220
+ | `resolve_package_versions` | Resolve exact versions via the real tool — npm/pnpm/yarn/pub/go/cargo/maven/gradle/pip, or Terraform/OpenTofu, Helm, and Ansible for infrastructure-shaped projects |
221
+
222
+ ---
223
+
224
+ ## Custom Skills
225
+
226
+ ```bash
227
+ nexus skill add my-code-review
228
+ # Edit .claude/commands/my-code-review.md
229
+ # Type /my-code-review in Claude Code
230
+ ```
231
+
232
+ Custom skills live alongside built-in skills in `.claude/commands/` — versioned in your repo, shared across the team.
233
+
234
+ ---
235
+
236
+ ## License
237
+
238
+ [MIT](LICENSE)