agentbrief 0.3.0 → 0.3.2
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/README.md
CHANGED
|
@@ -83,7 +83,7 @@ agentbrief use startup-founder # Grow: product, growth, launch strategy
|
|
|
83
83
|
| `growth-engineer` | Growth hacker with SEO audit + analytics + content strategy |
|
|
84
84
|
| `data-analyst` | BI analyst with metrics frameworks + SQL patterns |
|
|
85
85
|
| `startup-advisor` | Startup advisor with CEO review + launch planning |
|
|
86
|
-
| `social-media-manager` | Social media manager — Twitter/X + Xiaohongshu
|
|
86
|
+
| `social-media-manager` | Social media manager — Twitter/X + Xiaohongshu + Douyin research |
|
|
87
87
|
| `feishu-writer` | Feishu/Lark doc specialist — cloud docs, wikis, team notifications |
|
|
88
88
|
|
|
89
89
|
Browse the **[full Catalog](./CATALOG.md)** or the **[Website](https://0xranx.github.io/agentbrief)** for details on each brief.
|
package/README.zh-CN.md
CHANGED
|
@@ -83,7 +83,7 @@ agentbrief use startup-founder # 增长：产品、运营、发布策略
|
|
|
83
83
|
| `growth-engineer` | 增长工程师，SEO 审计 + 数据分析 + 内容策略 |
|
|
84
84
|
| `data-analyst` | 数据分析师，指标体系 + SQL + 数据叙事 |
|
|
85
85
|
| `startup-advisor` | 创业顾问，CEO 审查 + 发布规划 |
|
|
86
|
-
| `social-media-manager` | 社媒运营，Twitter/X +
|
|
86
|
+
| `social-media-manager` | 社媒运营，Twitter/X + 小红书 + 抖音调研 |
|
|
87
87
|
| `feishu-writer` | 飞书文档专家——云文档、知识库、团队通知 |
|
|
88
88
|
|
|
89
89
|
查看 **[完整目录](./CATALOG.md)** 或 **[官网](https://0xranx.github.io/agentbrief/index.zh.html)** 了解每个职能包的详情。
|
|
@@ -0,0 +1,65 @@
|
|
|
1
|
+
---
|
|
2
|
+
name: self-improving
|
|
3
|
+
description: >
|
|
4
|
+
Learn from corrections, mistakes, and discoveries during conversations.
|
|
5
|
+
When the user corrects you, when something fails unexpectedly, or when
|
|
6
|
+
you discover a project-specific pattern worth remembering — record it
|
|
7
|
+
as a learning so future sessions benefit automatically.
|
|
8
|
+
---
|
|
9
|
+
|
|
10
|
+
# Self-Improving
|
|
11
|
+
|
|
12
|
+
You have a persistent learning system. Use it to get smarter over time.
|
|
13
|
+
|
|
14
|
+
## When to Record a Learning
|
|
15
|
+
|
|
16
|
+
Record a learning when ANY of these happen:
|
|
17
|
+
|
|
18
|
+
1. **User corrects you** — "no, use X instead of Y", "don't do that", "that's wrong"
|
|
19
|
+
2. **Something fails unexpectedly** — a build error, test failure, or runtime crash reveals a project-specific gotcha
|
|
20
|
+
3. **You discover a non-obvious pattern** — the codebase has a convention that isn't documented anywhere
|
|
21
|
+
4. **User expresses a preference** — "I prefer X", "always do Y in this project", "never use Z"
|
|
22
|
+
|
|
23
|
+
Do NOT record:
|
|
24
|
+
- Generic programming knowledge (you already know this)
|
|
25
|
+
- Things already documented in CLAUDE.md or project docs
|
|
26
|
+
- Trivial one-time fixes
|
|
27
|
+
|
|
28
|
+
## How to Record
|
|
29
|
+
|
|
30
|
+
Create a markdown file in `.learnings/` at the project root:
|
|
31
|
+
|
|
32
|
+
```bash
|
|
33
|
+
# File naming: YYYY-MM-DD-short-description.md
|
|
34
|
+
.learnings/2026-03-20-use-pnpm-not-npm.md
|
|
35
|
+
```
|
|
36
|
+
|
|
37
|
+
Each learning file follows this format:
|
|
38
|
+
|
|
39
|
+
```markdown
|
|
40
|
+
# Use pnpm, not npm
|
|
41
|
+
|
|
42
|
+
**Context**: Ran `npm install` and got lockfile conflicts.
|
|
43
|
+
**Correction**: This project uses pnpm exclusively. The `pnpm-lock.yaml` is the source of truth.
|
|
44
|
+
**Rule**: Always use `pnpm` for install, add, and run commands. Never use `npm` or `yarn`.
|
|
45
|
+
```
|
|
46
|
+
|
|
47
|
+
Keep it short — 3-5 lines. One learning per file.
|
|
48
|
+
|
|
49
|
+
## How to Read Learnings
|
|
50
|
+
|
|
51
|
+
At the start of every conversation, check if `.learnings/` exists. If it does, read all files in it before starting work. These are hard-won lessons from previous sessions — respect them.
|
|
52
|
+
|
|
53
|
+
```
|
|
54
|
+
.learnings/
|
|
55
|
+
├── 2026-03-18-api-auth-requires-bearer.md
|
|
56
|
+
├── 2026-03-19-tests-need-env-setup.md
|
|
57
|
+
└── 2026-03-20-use-pnpm-not-npm.md
|
|
58
|
+
```
|
|
59
|
+
|
|
60
|
+
## Anti-patterns
|
|
61
|
+
|
|
62
|
+
- Don't create duplicate learnings — check existing files first
|
|
63
|
+
- Don't record learnings that contradict CLAUDE.md — CLAUDE.md wins
|
|
64
|
+
- Don't flood with trivial entries — quality over quantity
|
|
65
|
+
- Don't modify existing learnings unless the user explicitly corrects one
|
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
## Role
|
|
2
2
|
|
|
3
|
-
You are a social media manager who creates engaging content, manages posting workflows, and grows audience across platforms — Twitter/X
|
|
3
|
+
You are a social media manager who creates engaging content, manages posting workflows, and grows audience across platforms — Twitter/X, Xiaohongshu (小红书), and Douyin (抖音). You combine copywriting instincts with data-driven iteration: every post has a purpose, every thread tells a story, and every metric informs the next move.
|
|
4
4
|
|
|
5
5
|
## Tone
|
|
6
6
|
|
package/dist/cli.js
CHANGED
|
@@ -5,7 +5,7 @@ const program = new Command();
|
|
|
5
5
|
program
|
|
6
6
|
.name("agentbrief")
|
|
7
7
|
.description("One command turns your AI coding agent into a domain specialist")
|
|
8
|
-
.version("0.3.
|
|
8
|
+
.version("0.3.2");
|
|
9
9
|
program.addHelpText("after", `
|
|
10
10
|
Quick start:
|
|
11
11
|
agentbrief use ./path/to/brief Apply a brief to your project
|