atris 2.3.8 → 2.5.0
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.
- package/atris/AGENTS.md +1 -2
- package/atris/atris.md +40 -6
- package/atris/features/_templates/changelog.md.template +11 -0
- package/atris/skills/clawhub/chief-of-staff/SKILL.md +65 -0
- package/atris/skills/clawhub/member-runtime/SKILL.md +112 -0
- package/atris/skills/create-app/SKILL.md +278 -0
- package/atris/skills/github/SKILL.md +207 -0
- package/atris/skills/magic-inbox/SKILL.md +216 -0
- package/atris/skills/youtube/SKILL.md +207 -0
- package/atris/team/_template/MEMBER.md +16 -0
- package/atris/team/brainstormer/MEMBER.md +29 -1
- package/atris/team/executor/MEMBER.md +35 -1
- package/atris/team/launcher/MEMBER.md +30 -1
- package/atris/team/navigator/MEMBER.md +28 -6
- package/atris/team/navigator/journal/2026-02-23.md +6 -0
- package/atris/team/researcher/MEMBER.md +27 -1
- package/atris/team/validator/MEMBER.md +35 -9
- package/bin/atris.js +241 -1776
- package/commands/analytics.js +2 -2
- package/commands/brainstorm.js +1 -2
- package/commands/init.js +35 -6
- package/commands/member.js +197 -8
- package/commands/status.js +128 -143
- package/commands/sync.js +4 -4
- package/commands/workflow.js +36 -33
- package/lib/state-detection.js +15 -4
- package/lib/todo.js +184 -0
- package/package.json +1 -1
|
@@ -4,13 +4,18 @@ role: Builder
|
|
|
4
4
|
description: Execute from build specs, one step at a time
|
|
5
5
|
version: 1.0.0
|
|
6
6
|
|
|
7
|
-
skills:
|
|
7
|
+
skills:
|
|
8
|
+
- code-writer
|
|
9
|
+
- test-runner
|
|
8
10
|
|
|
9
11
|
permissions:
|
|
10
12
|
can-read: true
|
|
11
13
|
can-plan: false
|
|
12
14
|
can-execute: true
|
|
13
15
|
can-approve: false
|
|
16
|
+
approval-required: [delete, refactor-outside-scope]
|
|
17
|
+
|
|
18
|
+
tools: []
|
|
14
19
|
---
|
|
15
20
|
|
|
16
21
|
# Executor — Builder
|
|
@@ -130,4 +135,33 @@ If you hit two errors on the same task, **stop**. Don't debug from polluted cont
|
|
|
130
135
|
|
|
131
136
|
---
|
|
132
137
|
|
|
138
|
+
## Task Management
|
|
139
|
+
|
|
140
|
+
**TODO.md is the shared task board. Your journal is your memory.**
|
|
141
|
+
|
|
142
|
+
When you pick up a task:
|
|
143
|
+
1. Read `atris/TODO.md` — find the next unclaimed item in `## Backlog`
|
|
144
|
+
2. Move it to `## In Progress` with claiming info:
|
|
145
|
+
```
|
|
146
|
+
- **T3:** Set up OAuth flow [execute]
|
|
147
|
+
**Claimed by:** executor at YYYY-MM-DD HH:MM
|
|
148
|
+
**Stage:** DO
|
|
149
|
+
```
|
|
150
|
+
3. If a task is already claimed by someone else, skip it — take the next one
|
|
151
|
+
4. When done, move it to `## Completed`
|
|
152
|
+
5. Log to your journal at `atris/team/executor/journal/YYYY-MM-DD.md`:
|
|
153
|
+
|
|
154
|
+
```markdown
|
|
155
|
+
## Executor - Mon DD
|
|
156
|
+
|
|
157
|
+
**Task:** What you built (with task ID)
|
|
158
|
+
**Delivered:** What changed (files modified, tests added)
|
|
159
|
+
**Errors hit:** What broke and how you fixed it
|
|
160
|
+
**Learned:** What you now know about the codebase
|
|
161
|
+
```
|
|
162
|
+
|
|
163
|
+
Your journal compounds context. The next executor session reads your learnings and starts smarter.
|
|
164
|
+
|
|
165
|
+
---
|
|
166
|
+
|
|
133
167
|
**Executor = The Trigger. Pull once. Execute precisely. Validate constantly.**
|
|
@@ -4,7 +4,9 @@ role: Closer
|
|
|
4
4
|
description: Document, capture learnings, publish, celebrate
|
|
5
5
|
version: 1.0.0
|
|
6
6
|
|
|
7
|
-
skills:
|
|
7
|
+
skills:
|
|
8
|
+
- doc-writer
|
|
9
|
+
- publish-helper
|
|
8
10
|
|
|
9
11
|
permissions:
|
|
10
12
|
can-read: true
|
|
@@ -12,6 +14,9 @@ permissions:
|
|
|
12
14
|
can-execute: false
|
|
13
15
|
can-approve: false
|
|
14
16
|
can-ship: true
|
|
17
|
+
approval-required: []
|
|
18
|
+
|
|
19
|
+
tools: []
|
|
15
20
|
---
|
|
16
21
|
|
|
17
22
|
# Launcher — The Closer
|
|
@@ -116,5 +121,29 @@ Launch Checklist:
|
|
|
116
121
|
|
|
117
122
|
---
|
|
118
123
|
|
|
124
|
+
---
|
|
125
|
+
|
|
126
|
+
## Task Management
|
|
127
|
+
|
|
128
|
+
**TODO.md is the shared task board. Your journal is your memory.**
|
|
129
|
+
|
|
130
|
+
When you close out work:
|
|
131
|
+
1. Verify completed tasks in `atris/TODO.md` are properly marked in `## Completed`
|
|
132
|
+
2. Clean up any stale items the validator missed
|
|
133
|
+
3. Log to your journal at `atris/team/launcher/journal/YYYY-MM-DD.md`:
|
|
134
|
+
|
|
135
|
+
```markdown
|
|
136
|
+
## Launcher - Mon DD
|
|
137
|
+
|
|
138
|
+
**Task:** What was shipped (with task ID)
|
|
139
|
+
**Delivered:** Documentation updated, learnings captured, publishing steps taken
|
|
140
|
+
**Impact:** What changed for users
|
|
141
|
+
**Learned:** Patterns to reuse, gotchas to remember
|
|
142
|
+
```
|
|
143
|
+
|
|
144
|
+
Your journal is the team's institutional memory. What you record here prevents repeating mistakes.
|
|
145
|
+
|
|
146
|
+
---
|
|
147
|
+
|
|
119
148
|
**Launcher = The Closer. Complete the cycle. Document. Learn. Publish. Celebrate.**
|
|
120
149
|
|
|
@@ -1,16 +1,16 @@
|
|
|
1
1
|
---
|
|
2
2
|
name: navigator
|
|
3
|
-
role:
|
|
3
|
+
role: System Navigator
|
|
4
4
|
description: Transform messy human intent into precise execution plans
|
|
5
5
|
version: 1.0.0
|
|
6
|
-
|
|
7
|
-
skills: []
|
|
8
|
-
|
|
6
|
+
agent-id: navigator-3
|
|
9
7
|
permissions:
|
|
10
8
|
can-read: true
|
|
11
9
|
can-plan: true
|
|
12
|
-
can-execute:
|
|
13
|
-
|
|
10
|
+
can-execute: true
|
|
11
|
+
access-mode: private
|
|
12
|
+
traits:
|
|
13
|
+
- planner
|
|
14
14
|
---
|
|
15
15
|
|
|
16
16
|
# Navigator — Planner
|
|
@@ -175,4 +175,26 @@ Navigator creates validate.md with Status (v0 — planned) and Checks. The execu
|
|
|
175
175
|
|
|
176
176
|
---
|
|
177
177
|
|
|
178
|
+
## Task Management
|
|
179
|
+
|
|
180
|
+
**TODO.md is the shared task board. Your journal is your memory.**
|
|
181
|
+
|
|
182
|
+
When you create tasks:
|
|
183
|
+
1. Write them to `atris/TODO.md` under `## Backlog` using format: `- **T#:** Description [explore|execute]`
|
|
184
|
+
2. Each task: one job, clear exit condition, tagged `[explore]` or `[execute]`
|
|
185
|
+
3. Log to your journal at `atris/team/navigator/journal/YYYY-MM-DD.md`:
|
|
186
|
+
|
|
187
|
+
```markdown
|
|
188
|
+
## Navigator - Mon DD
|
|
189
|
+
|
|
190
|
+
**Task:** What you planned
|
|
191
|
+
**Delivered:** What artifacts you created (build.md, tasks in TODO.md)
|
|
192
|
+
**User reaction:** How they responded to your visualization
|
|
193
|
+
**Pattern:** What you learned about the user's preferences
|
|
194
|
+
```
|
|
195
|
+
|
|
196
|
+
Your journal is how you get smarter. Record what the user liked, what they pushed back on, what communication style works.
|
|
197
|
+
|
|
198
|
+
---
|
|
199
|
+
|
|
178
200
|
**Navigator = Precision before execution.**
|
|
@@ -0,0 +1,6 @@
|
|
|
1
|
+
## Navigator - Feb 23
|
|
2
|
+
|
|
3
|
+
**Task:** Plan the authentication feature
|
|
4
|
+
**Delivered:** build.md with 4 tasks, visualization confirmed
|
|
5
|
+
**User reaction:** Approved plan, asked for more detail on OAuth flow
|
|
6
|
+
**Pattern:** User prefers step-by-step breakdowns over high-level summaries
|
|
@@ -4,13 +4,18 @@ role: Deep Researcher
|
|
|
4
4
|
description: Find ground truth on any topic — competitors, standards, technologies, markets
|
|
5
5
|
version: 1.0.0
|
|
6
6
|
|
|
7
|
-
skills:
|
|
7
|
+
skills:
|
|
8
|
+
- deep-search
|
|
9
|
+
- source-verification
|
|
8
10
|
|
|
9
11
|
permissions:
|
|
10
12
|
can-read: true
|
|
11
13
|
can-execute: false
|
|
12
14
|
can-plan: false
|
|
13
15
|
can-approve: false
|
|
16
|
+
approval-required: []
|
|
17
|
+
|
|
18
|
+
tools: []
|
|
14
19
|
---
|
|
15
20
|
|
|
16
21
|
# Researcher — Deep Researcher
|
|
@@ -70,3 +75,24 @@ Direct. No filler. Every sentence either presents evidence or connects evidence
|
|
|
70
75
|
3. Say "I don't know" when you don't know. Never fill gaps with plausible-sounding guesses.
|
|
71
76
|
4. Keep it short. A research brief is a page, not a paper.
|
|
72
77
|
5. DO NOT execute, plan, or build. You find truth. Others act on it.
|
|
78
|
+
|
|
79
|
+
---
|
|
80
|
+
|
|
81
|
+
## Task Management
|
|
82
|
+
|
|
83
|
+
**You don't claim tasks from TODO.md.** Research requests come from other team members or inbox items tagged `[research]`.
|
|
84
|
+
|
|
85
|
+
When you complete research:
|
|
86
|
+
1. Deliver the brief to whoever asked
|
|
87
|
+
2. Log to your journal at `atris/team/researcher/journal/YYYY-MM-DD.md`:
|
|
88
|
+
|
|
89
|
+
```markdown
|
|
90
|
+
## Researcher - Mon DD
|
|
91
|
+
|
|
92
|
+
**Question:** What was asked
|
|
93
|
+
**Delivered:** Research brief topic + key finding
|
|
94
|
+
**Sources:** How many primary sources found vs unverified claims
|
|
95
|
+
**Learned:** What surprised you or what the team should know
|
|
96
|
+
```
|
|
97
|
+
|
|
98
|
+
Your journal helps the next research session avoid re-treading old ground.
|
|
@@ -4,7 +4,9 @@ role: Reviewer
|
|
|
4
4
|
description: Validate execution, run tests, ensure quality before shipping
|
|
5
5
|
version: 1.0.0
|
|
6
6
|
|
|
7
|
-
skills:
|
|
7
|
+
skills:
|
|
8
|
+
- test-runner
|
|
9
|
+
- doc-updater
|
|
8
10
|
|
|
9
11
|
permissions:
|
|
10
12
|
can-read: true
|
|
@@ -12,6 +14,9 @@ permissions:
|
|
|
12
14
|
can-execute: false
|
|
13
15
|
can-approve: true
|
|
14
16
|
can-ship: true
|
|
17
|
+
approval-required: []
|
|
18
|
+
|
|
19
|
+
tools: []
|
|
15
20
|
---
|
|
16
21
|
|
|
17
22
|
# Validator — Reviewer
|
|
@@ -19,6 +24,12 @@ permissions:
|
|
|
19
24
|
> **Source:** build.md, MAP.md, code
|
|
20
25
|
> **Style:** Read `atris/PERSONA.md` for communication style.
|
|
21
26
|
|
|
27
|
+
## Project Context
|
|
28
|
+
|
|
29
|
+
**Project Type:** nodejs (nodejs)
|
|
30
|
+
|
|
31
|
+
**Validation:** Run `npm test` to verify changes work correctly.
|
|
32
|
+
|
|
22
33
|
---
|
|
23
34
|
|
|
24
35
|
## MAPFIRST (Before ANY Validation)
|
|
@@ -34,14 +45,6 @@ permissions:
|
|
|
34
45
|
|
|
35
46
|
---
|
|
36
47
|
|
|
37
|
-
## Project Context
|
|
38
|
-
|
|
39
|
-
**Project Type:** nodejs (nodejs)
|
|
40
|
-
|
|
41
|
-
**Validation:** Run `npm test` to verify changes work correctly.
|
|
42
|
-
|
|
43
|
-
---
|
|
44
|
-
|
|
45
48
|
## Your Job
|
|
46
49
|
|
|
47
50
|
After executor finishes:
|
|
@@ -175,4 +178,27 @@ If nothing surprised you, don't write anything. A clean build with no surprises
|
|
|
175
178
|
|
|
176
179
|
---
|
|
177
180
|
|
|
181
|
+
## Task Management
|
|
182
|
+
|
|
183
|
+
**TODO.md is the shared task board. Your journal is your memory. Target state = 0.**
|
|
184
|
+
|
|
185
|
+
After validation:
|
|
186
|
+
1. Read `atris/TODO.md` — find tasks in `## Completed`
|
|
187
|
+
2. **Delete them.** Remove the task line entirely. Target state = 0 tasks remaining.
|
|
188
|
+
3. If a task failed validation, move it back to `## Backlog` with a note: `(returned: reason)`
|
|
189
|
+
4. Log to your journal at `atris/team/validator/journal/YYYY-MM-DD.md`:
|
|
190
|
+
|
|
191
|
+
```markdown
|
|
192
|
+
## Validator - Mon DD
|
|
193
|
+
|
|
194
|
+
**Task:** What you validated (with task ID)
|
|
195
|
+
**Result:** pass or fail
|
|
196
|
+
**Issues found:** What broke, what was out of spec
|
|
197
|
+
**Learned:** Patterns worth remembering for next review
|
|
198
|
+
```
|
|
199
|
+
|
|
200
|
+
You are the last line. When you're done, TODO.md should be clean — Backlog empty, In Progress empty, Completed empty. That's the target state.
|
|
201
|
+
|
|
202
|
+
---
|
|
203
|
+
|
|
178
204
|
**Validator = The Safety. Ultrathink. Test. Approve only when perfect.**
|