@kendoo.agentdesk/agentdesk 0.8.0 → 0.8.1

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 (2) hide show
  1. package/README.md +15 -4
  2. package/package.json +1 -1
package/README.md CHANGED
@@ -79,6 +79,7 @@ agentdesk update Update to the latest version
79
79
  |------|-------------|
80
80
  | `--description`, `-d` | Task description or requirements |
81
81
  | `--cwd` | Working directory (defaults to current) |
82
+ | `--legacy` | Use single-process mode (cheaper, less independent reasoning) |
82
83
 
83
84
  ## Configuration
84
85
 
@@ -115,12 +116,22 @@ AgentDesk also auto-discovers agents from `.claude/agents/`, `.claude/commands/`
115
116
 
116
117
  ## How It Works
117
118
 
119
+ Each agent runs as its own independent Claude session — not one model role-playing multiple personas. This produces genuine independent reasoning, real disagreements, and better output.
120
+
118
121
  1. You run `agentdesk team TASK-ID` in your project directory
119
122
  2. AgentDesk detects your project type, reads `CLAUDE.md`, and discovers existing agents
120
- 3. The team collaborates in structured phases: **Intake** > **Brainstorm** > **Planning** > **Execution** > **Review**
121
- 4. The session streams live to [agentdesk.live](https://agentdesk.live) where you can watch and send messages to the team
122
- 5. For UI tasks, the team captures screenshots (desktop + mobile) and uploads them to your task tracker
123
- 6. Token usage is tracked and displayed per session
123
+ 3. The orchestrator manages 5 phases: **Intake** > **Brainstorm** > **Planning** > **Execution** > **Review**
124
+ 4. In brainstorm/planning/review, all agents run **in parallel** each gets their own Claude session
125
+ 5. In execution, agents run sequentially: Dennis implements > Sam audits > Luna+Mark review > Vera tests > Bart QA
126
+ 6. The session streams live to [agentdesk.live](https://agentdesk.live) where you can watch and send messages to the team
127
+ 7. Token usage is tracked and displayed per session
128
+
129
+ ### Session Modes
130
+
131
+ | Mode | Command | Description |
132
+ |------|---------|-------------|
133
+ | Sub-agents (default) | `agentdesk team -d "..."` | Each agent is an independent Claude session. Better reasoning, ~4-5x tokens. |
134
+ | Legacy | `agentdesk team -d "..." --legacy` | Single Claude process, all agents as personas. Cheaper, less independent. |
124
135
 
125
136
  ## Daemon (Remote Sessions)
126
137
 
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@kendoo.agentdesk/agentdesk",
3
- "version": "0.8.0",
3
+ "version": "0.8.1",
4
4
  "description": "AI team orchestrator for Claude Code — run collaborative agent sessions from your terminal",
5
5
  "type": "module",
6
6
  "bin": {