@evolvehq/docflow 0.6.0 → 0.7.0

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 (3) hide show
  1. package/README.md +53 -5
  2. package/USAGE.md +27 -6
  3. package/package.json +1 -1
package/README.md CHANGED
@@ -2,8 +2,9 @@
2
2
 
3
3
  ![docflow — ADR-driven documentation workflow](docs/preview.png)
4
4
 
5
- A plugin for **ADR-driven, documentation-led projects**, working on both
6
- the **pi** and **Claude Code** coding agents from the same skill files.
5
+ A plugin for **ADR-driven, documentation-led projects**, working on
6
+ **Claude Code, Claude Cowork, pi, Codex, and OpenCode** from the same
7
+ skill files (see [Install](#install)).
7
8
  It installs a `bootstrap` skill that scaffolds (or retrofits) an
8
9
  **Architecture Decision Record (ADR)** catalogue, a plan queue, and
9
10
  `AGENTS.md` conventions into any repository, plus a set of **lifecycle
@@ -66,9 +67,23 @@ rather than overwriting them).
66
67
 
67
68
  ## Install
68
69
 
69
- docflow ships for two coding agents from the **same** skill files — only
70
- the manifest differs (`.claude-plugin/` for Claude Code, `package.json`
71
- for pi).
70
+ docflow ships from **one `skills/` tree** to five coding agents — only
71
+ the packaging differs. Two surfaces: the scaffolded **output**
72
+ (`AGENTS.md`, the ADR catalogue, `plan/`, `_agent/`) is plain Markdown
73
+ read natively by any agent that loads `AGENTS.md`; the **skills** are
74
+ `SKILL.md` files the host discovers.
75
+
76
+ | Agent | Output | Skills | Install | Invoke |
77
+ |-------|:------:|:------:|---------|--------|
78
+ | Claude Code | native | ✅ | marketplace (below) | `/bootstrap` |
79
+ | Claude Cowork | native | ✅ | same Claude Code plugin | `/bootstrap` |
80
+ | pi | native | ✅ | `pi install npm:@evolvehq/docflow` | `/skill:bootstrap` |
81
+ | Codex | native | ✅ | copy into `~/.agents/skills/` | `$bootstrap` / `/skills` |
82
+ | OpenCode | native | ✅ | reads `.claude`/`.agents`/`.opencode` skills | auto, by description |
83
+
84
+ Handy: `~/.agents/skills/` is read by **both Codex and OpenCode**, and
85
+ `~/.claude/skills/` by **both Claude Code and OpenCode** — so one install
86
+ often serves two agents.
72
87
 
73
88
  ### Claude Code — from this marketplace
74
89
 
@@ -80,6 +95,13 @@ for pi).
80
95
  Invoke with `/bootstrap`, `/new-adr`, `/ship-item`, … (auto-triggers on
81
96
  matching requests too).
82
97
 
98
+ ### Claude Cowork
99
+
100
+ Cowork uses the **same plugin system** as Claude Code, so install the
101
+ docflow plugin exactly as above (`/plugin marketplace add
102
+ EvolveHQ/docflow`, then install) — or from Anthropic's community
103
+ marketplace once listed. No separate packaging.
104
+
83
105
  ### pi coding agent
84
106
 
85
107
  ```
@@ -97,6 +119,32 @@ The scaffolded output (`AGENTS.md`, `CONVENTIONS.md`, the ADR catalogue,
97
119
  `plan/`, `_agent/`) is plain Markdown and is read natively by pi's
98
120
  hierarchical `AGENTS.md` loading — no porting needed.
99
121
 
122
+ ### Codex (OpenAI)
123
+
124
+ Codex reads `AGENTS.md` natively and discovers skills from `.agents/skills`.
125
+ Copy docflow's skills into a path Codex scans:
126
+
127
+ ```
128
+ # user-wide (all projects), from a clone of this repo:
129
+ mkdir -p ~/.agents/skills && cp -r <docflow>/skills/* ~/.agents/skills/
130
+ # or from npm:
131
+ npm install @evolvehq/docflow
132
+ cp -r node_modules/@evolvehq/docflow/skills/* ~/.agents/skills/
133
+ ```
134
+
135
+ On Windows, copy `skills\*` into `%USERPROFILE%\.agents\skills\`. Invoke
136
+ with `$bootstrap` / `/skills`, or just describe the task (Codex
137
+ auto-triggers from the skill description). The structured assessment
138
+ questions fall back to plain `A/B/C` text where there is no select tool.
139
+
140
+ ### OpenCode (sst)
141
+
142
+ OpenCode scans `.opencode/skills/`, **`.claude/skills/`, and
143
+ `.agents/skills/`** (project and global), so a Claude Code or Codex
144
+ install is **picked up automatically** — or copy `skills/*` into
145
+ `~/.config/opencode/skills/` (or per-project `.opencode/skills/`). Skills
146
+ auto-load by description.
147
+
100
148
  ### Claude Code — local development (no install)
101
149
 
102
150
  ```
package/USAGE.md CHANGED
@@ -164,12 +164,33 @@ The repo is now ready for ADR-driven work. Typical first steps:
164
164
  ## 5a. Lifecycle skills
165
165
 
166
166
  The manual workflow in §5 is exactly what the lifecycle skills
167
- automate. They all share three properties: they **read `CONVENTIONS.md`
168
- first** and honour the repo's recorded choices (ADR shape, status
169
- lifecycle, integration model, multi-agent mode); they **refuse on an
170
- un-bootstrapped repo** and point at `/bootstrap`; and they keep
171
- `INDEX.md` and the `_agent/` coordination files in sync as a side
172
- effect.
167
+ automate. They all share four properties: they **run a short assessment
168
+ first** (see below); they **read `CONVENTIONS.md` first** and honour the
169
+ repo's recorded choices (ADR shape, status lifecycle, integration model,
170
+ multi-agent mode); they **refuse on an un-bootstrapped repo** and point
171
+ at `/bootstrap`; and they keep `INDEX.md` and the `_agent/` coordination
172
+ files in sync as a side effect.
173
+
174
+ **The assessment protocol.** `new-adr`, `new-plan`, `add-convention`,
175
+ `brainstorm`, and `agent-wave` each open with a brief assessment before
176
+ acting — the same pattern bootstrap uses (§3):
177
+
178
+ - An **opt-out gate** first: *run the assessment, or skip it?* The
179
+ recommended default flips on context — *run* when you invoked the skill
180
+ with little detail, *skip* when your request already specifies
181
+ everything.
182
+ - Questions are asked **one at a time**, each with a **recommended
183
+ option** you can accept, override, or replace.
184
+ - Answers use **scroll-to-select** (single- or multiple-choice) where the
185
+ host exposes a selection tool (Claude Code's `AskUserQuestion`), falling
186
+ back to plain `A/B/C` lists otherwise. Free text is used only where an
187
+ enumerable set is impossible (e.g. an ADR title).
188
+ - **You decide** — a skill never proceeds past a question without your
189
+ input, and never guesses scope when invoked with no context.
190
+
191
+ `agent-wave` additionally asks the wave parallelism and the merge /
192
+ integration strategy. So a bare *"new ADR"* leads with a couple of quick
193
+ picks; a fully-specified request lets you skip straight through.
173
194
 
174
195
  | Skill | Replaces these manual steps |
175
196
  |-------|------------------------------|
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@evolvehq/docflow",
3
- "version": "0.6.0",
3
+ "version": "0.7.0",
4
4
  "description": "ADR-driven documentation workflow: scaffold an Architecture Decision Record (ADR) catalogue, a plan/ queue, and AGENTS.md conventions into any repo, then author, queue, ship, and audit ADRs with lifecycle skills. Works with the pi coding agent and Claude Code.",
5
5
  "keywords": [
6
6
  "pi-package",