clickup-agent-cli 0.3.0 → 0.4.1
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.
- package/.claude-plugin/marketplace.json +1 -1
- package/.claude-plugin/plugin.json +1 -1
- package/README.md +33 -10
- package/dist/clickup.js +474 -293
- package/package.json +11 -3
- package/skills/clickup/SKILL.md +7 -0
- package/skills/clickup-custom-report/SKILL.md +1 -1
- package/skills/clickup-fields/SKILL.md +3 -3
- package/skills/clickup-rollout/SKILL.md +2 -2
- package/skills/clickup-task-triage/SKILL.md +1 -1
- package/skills/clickup-tasks/SKILL.md +10 -4
- package/skills/clickup-time/SKILL.md +3 -2
|
@@ -11,7 +11,7 @@
|
|
|
11
11
|
"name": "clickup",
|
|
12
12
|
"source": "./",
|
|
13
13
|
"description": "ClickUp CLI with 25 agent skills covering the full API -- token-efficient alternative to MCP with chat, time tracking, docs, and project management workflows",
|
|
14
|
-
"version": "0.
|
|
14
|
+
"version": "0.4.1",
|
|
15
15
|
"homepage": "https://github.com/henryreith/clickup-cli",
|
|
16
16
|
"keywords": ["clickup", "project-management", "tasks", "time-tracking"],
|
|
17
17
|
"category": "productivity"
|
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "clickup",
|
|
3
3
|
"description": "ClickUp CLI with 25 agent skills covering the full API -- token-efficient alternative to MCP with chat, time tracking, docs, and project management workflows",
|
|
4
|
-
"version": "0.
|
|
4
|
+
"version": "0.4.1",
|
|
5
5
|
"author": {
|
|
6
6
|
"name": "Henry Reith"
|
|
7
7
|
},
|
package/README.md
CHANGED
|
@@ -80,7 +80,7 @@ Every command supports multiple output formats:
|
|
|
80
80
|
|
|
81
81
|
## AI Agent Usage
|
|
82
82
|
|
|
83
|
-
The CLI is built from the ground up for AI agents, following the "CLI as execution layer, skills as guidance layer" pattern. It ships as a Claude Code plugin with
|
|
83
|
+
The CLI is built from the ground up for AI agents, following the "CLI as execution layer, skills as guidance layer" pattern. It ships as a Claude Code plugin with 25 agent skills, and works with any agent platform that can run bash commands.
|
|
84
84
|
|
|
85
85
|
**vs ClickUp MCP:** This CLI covers the same operations as the official ClickUp MCP server (tasks, docs, time tracking, chat, members, hierarchy) plus more, while consuming far fewer tokens. A full skills hierarchy loads in ~150 tokens at session start vs injecting an entire API schema. Works in any agent platform that supports bash -- not just MCP-compatible hosts.
|
|
86
86
|
|
|
@@ -93,14 +93,14 @@ Install the ClickUp CLI as a Claude Code plugin for zero-friction access to all
|
|
|
93
93
|
/plugin marketplace add henryreith/clickup-cli
|
|
94
94
|
|
|
95
95
|
# Install the plugin
|
|
96
|
-
/plugin install clickup@clickup-cli
|
|
96
|
+
/plugin install clickup@clickup-agent-cli
|
|
97
97
|
|
|
98
98
|
# Use skills directly
|
|
99
99
|
/clickup:weekly-review 12345678
|
|
100
100
|
/clickup:sprint-planning list-id-here
|
|
101
101
|
```
|
|
102
102
|
|
|
103
|
-
Once installed, Claude Code auto-discovers all
|
|
103
|
+
Once installed, Claude Code auto-discovers all 25 skills and loads them on demand. Recipe skills like `/clickup:weekly-review` run in isolated subagents with full ClickUp CLI access.
|
|
104
104
|
|
|
105
105
|
### Claude Agent SDK
|
|
106
106
|
|
|
@@ -112,7 +112,7 @@ import { query } from "@anthropic-ai/claude-agent-sdk";
|
|
|
112
112
|
for await (const message of query({
|
|
113
113
|
prompt: "Create a task for the login bug fix and assign to the backend team",
|
|
114
114
|
options: {
|
|
115
|
-
plugins: [{ type: "local", path: "./node_modules/clickup-cli" }],
|
|
115
|
+
plugins: [{ type: "local", path: "./node_modules/clickup-agent-cli" }],
|
|
116
116
|
allowedTools: ["Skill", "Bash"],
|
|
117
117
|
settingSources: ["project"]
|
|
118
118
|
}
|
|
@@ -133,7 +133,7 @@ clickup skill show clickup
|
|
|
133
133
|
clickup skill show clickup-tasks
|
|
134
134
|
|
|
135
135
|
# Or asks the CLI what fields are needed (~50 tokens in response)
|
|
136
|
-
clickup schema
|
|
136
|
+
clickup schema task.create
|
|
137
137
|
|
|
138
138
|
# For complex workflows, load a recipe skill
|
|
139
139
|
clickup skill show clickup-weekly-review
|
|
@@ -141,8 +141,8 @@ clickup skill show clickup-weekly-review
|
|
|
141
141
|
|
|
142
142
|
**Three skill tiers:**
|
|
143
143
|
- **Root skill** - Index and router. What the CLI does, how to discover more.
|
|
144
|
-
- **Sub-skills** (
|
|
145
|
-
- **Recipe skills** (
|
|
144
|
+
- **Sub-skills** (11) - Per-resource command reference. Tasks, spaces, comments, time tracking, chat, etc.
|
|
145
|
+
- **Recipe skills** (13) - Multi-step workflow guides that accept natural language arguments. Scope any recipe to a specific team, department, or person.
|
|
146
146
|
|
|
147
147
|
**Example recipe invocations:**
|
|
148
148
|
```bash
|
|
@@ -180,11 +180,34 @@ This creates `/marketing-weekly` alongside the built-in `/clickup:*` skills.
|
|
|
180
180
|
| Platform | How to use |
|
|
181
181
|
|----------|-----------|
|
|
182
182
|
| Claude Code | Plugin via marketplace (see above) |
|
|
183
|
-
| Claude Agent SDK | `plugins: [{ type: "local", path: "./node_modules/clickup-cli" }]` |
|
|
183
|
+
| Claude Agent SDK | `plugins: [{ type: "local", path: "./node_modules/clickup-agent-cli" }]` |
|
|
184
184
|
| Gemini CLI | `npm install -g clickup-agent-cli`, read skills via `clickup skill show` |
|
|
185
185
|
| OpenAI Codex | `npm install -g clickup-agent-cli`, execute commands via bash |
|
|
186
186
|
| Custom agents | `clickup skill list` for discovery, `clickup schema` for field info |
|
|
187
187
|
|
|
188
|
+
### Bootstrap From Any Agent
|
|
189
|
+
|
|
190
|
+
Any agent that can run shell commands can set up and use the CLI in five steps, no plugin system required:
|
|
191
|
+
|
|
192
|
+
```bash
|
|
193
|
+
# 1. Install
|
|
194
|
+
npm install -g clickup-agent-cli
|
|
195
|
+
|
|
196
|
+
# 2. Authenticate non-interactively (env var, or --token-file / --token per call)
|
|
197
|
+
export CLICKUP_API_TOKEN=pk_your_token
|
|
198
|
+
|
|
199
|
+
# 3. Verify auth and workspace
|
|
200
|
+
clickup config validate
|
|
201
|
+
|
|
202
|
+
# 4. Discover capabilities: the root skill indexes everything
|
|
203
|
+
clickup skill show clickup # start here (~150 tokens)
|
|
204
|
+
clickup skill list # all 25 skills
|
|
205
|
+
clickup skill show clickup-tasks # per-resource command reference
|
|
206
|
+
|
|
207
|
+
# 5. Discover exact fields before calling a command
|
|
208
|
+
clickup schema task.create
|
|
209
|
+
```
|
|
210
|
+
|
|
188
211
|
### Agent-Friendly Design
|
|
189
212
|
|
|
190
213
|
- JSON is the default output when stdout is not a TTY
|
|
@@ -199,10 +222,10 @@ This creates `/marketing-weekly` alongside the built-in `/clickup:*` skills.
|
|
|
199
222
|
|
|
200
223
|
```bash
|
|
201
224
|
# Discover what's available
|
|
202
|
-
clickup schema
|
|
225
|
+
clickup schema task
|
|
203
226
|
|
|
204
227
|
# Check what fields task create needs
|
|
205
|
-
clickup schema
|
|
228
|
+
clickup schema task.create --format json
|
|
206
229
|
|
|
207
230
|
# Create a task (JSON output automatic in non-TTY)
|
|
208
231
|
clickup task create --list-id 998877 --name "Review PR" --priority 2
|