菜单
Log learnings, errors, and feature requests for continuous improvement. Use when: (1) a command fails, (2) user corrects you, (3) user requests something you can't do, (4) you discover outdated knowledge, (5) you find a better approach to a recurring task.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Email deliverability, list management, sequences, segmentation, and campaign optimization.
Daily macro + market brief (FRED + benchmarks + watchlist ticker) with critical-headline triage, explicit source/freshness notes, and graceful fallback behavior. Use when the user asks for a concise “what moved today and why” summary with practical context.
Write content that reads as naturally human — no AI tells, no corporate fluff. Use when drafting, editing, or reviewing any content meant for publication or external audiences: blog posts, articles, newsletters, marketing copy, social media threads, client-facing documents, website copy, emails to lists. NOT for: casual chat, code, internal notes, or quick replies.
Daily journaling through your agent. Smart prompts, mood tracking, weekly reflections, pattern detection. Private, local, no cloud.
Research markets with sizing, segmentation, competitor mapping, pricing checks, and demand validation that turn fuzzy ideas into decision-ready evidence. Use when (1) you need TAM, SAM, SOM, whitespace, or category sizing; (2) you must compare competitors, pricing, positioning, or customer segments before acting; (3) the user asks whether a niche, launch, expansion, or go-to-market bet is actually worth pursuing.
Creates a morning briefing with priorities, calendar, and key updates
| name | self-improvement |
| description | Log learnings, errors, and feature requests for continuous improvement. Use when: (1) a command fails, (2) user corrects you, (3) user requests something you can't do, (4) you discover outdated knowledge, (5) you find a better approach to a recurring task. |
Log mistakes, corrections, and discoveries to .learnings/ files so you get smarter over time.
mkdir -p ~/workspace/.learnings
| Situation | Action |
|---|---|
| Command fails | Log to .learnings/ERRORS.md |
| User corrects you | Log to .learnings/LEARNINGS.md |
| User wants missing feature | Log to .learnings/FEATURE_REQUESTS.md |
| Found better approach | Log to .learnings/LEARNINGS.md |
| Knowledge was outdated | Log to .learnings/LEARNINGS.md |
When a learning proves broadly useful, promote it to a permanent file:
| Learning Type | Promote To |
|---|---|
| Behavioral patterns | SOUL.md |
| Workflow improvements | AGENTS.md |
| Tool gotchas | TOOLS.md |
Append to .learnings/LEARNINGS.md:
## [LRN-YYYYMMDD-001] category
**Logged**: 2026-03-05
**Priority**: low | medium | high | critical
**Status**: pending
### Summary
One-line description
### Details
What happened, what was wrong, what's correct
### Suggested Action
Specific fix or improvement
Append to .learnings/ERRORS.md:
## [ERR-YYYYMMDD-001] what_failed
**Logged**: 2026-03-05
**Priority**: high
**Status**: pending
### Summary
What failed
### Error
Actual error message
### Context
What you were trying to do
### Suggested Fix
How to prevent this next time
Append to .learnings/FEATURE_REQUESTS.md:
## [FEAT-YYYYMMDD-001] capability_name
**Logged**: 2026-03-05
**Priority**: medium
**Status**: pending
### Requested Capability
What the user wanted
### Suggested Implementation
How this could work
Log automatically when you notice:
When fixed, update the entry:
**Status**: pending → **Status**: resolvedCheck your learnings periodically:
# Count pending items
grep -c "Status.*pending" ~/workspace/.learnings/*.md
# Find high priority items
grep -B3 "Priority.*high" ~/workspace/.learnings/*.md