菜单
This skill should be used when the user invokes "/personal:dev" with a subcommand. Subcommands: init (project + task setup), new (create task from description or GH issue), complete (finish a task), renumber (reorder task priorities), prime (load project context).
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
This skill should be used when the user invokes "/personal:morning" to start their morning planning session. Provides briefing, task overview, and guided day planning conversation.
Danny's writing and editing guide. Use when writing, editing, or reviewing text of any kind. Covers general writing principles, Danny's distinctive voice, style rules, and anti-slop patterns. Use for blog posts, docs, emails, proposals, personal essays, knowledge base articles, GitHub issues, or any other non-trivial text intended for others to read. Also use for quality checking and deslopping.
Remove AI slop and corporate bullshit from text — produces clean, natural, human-sounding writing. Use when the user asks to "unslopify", "deslop", "remove slop", "clean up AI text", or wants text cleaned of AI-sounding language without applying Danny's specific voice.
Rewrite text in Danny's distinctive voice — conversational, substantive, cuts through bullshit. Use when the user asks to "danify", "rewrite in my voice", "make this sound like me", or wants text transformed into Danny's writing style.
Comprehensive writing review — spawns analyser agents for thorough quality and voice assessment. Use when the user asks for a "full check", "thorough review", "comprehensive edit", or wants detailed quality analysis of their writing.
Quick writing quality check — scans for slop, style issues, spelling/grammar errors. Use when the user asks for a "quick check", "quick review", "scan for issues", or wants a fast quality pass without full voice analysis.
| name | dev |
| description | This skill should be used when the user invokes "/personal:dev" with a subcommand. Subcommands: init (project + task setup), new (create task from description or GH issue), complete (finish a task), renumber (reorder task priorities), prime (load project context). |
| argument-hint | <init|new|complete|renumber|prime> [args] |
| disable-model-invocation | true |
| allowed-tools | ["Bash","Read","Write","Edit","Glob","Grep"] |
Project development utilities for initialisation, task management, and context priming.
Designed for any dev project. Tasks are tracked as markdown files: todo tasks in
docs/tasks-todo/task-NUMBER-name.md, completed tasks in
docs/tasks-done/YYYY-MM-DD-task-NUMBER-name.md. If this project has not been initialised
with this system yet, suggest the user runs /personal:dev init first.
$ARGUMENTS
Scripts in scripts/ are executable with shebangs. Execute them directly — never
prepend bash, sh, or any interpreter. For example: scripts/init-project.sh not
bash scripts/init-project.sh.
The first word of the user input above is the subcommand. Everything after it is context for that subcommand. If no argument was provided, list the available subcommands and ask the user which they want.
Initialise or update a project's AI boilerplate and task management structure.
Two phases:
scripts/init-project.shCreate a new task file in docs/tasks-todo/.
Read and follow references/new.md. The context for the new task is everything after "new" in the user input above.
Complete a task by moving it from docs/tasks-todo/ to docs/tasks-done/ with a date prefix.
The task identifier is the word immediately after "complete" in the user input.
Execute: scripts/complete-task.sh <task-identifier>
Report the result to the user.
Renumber prioritised tasks to start from 1 with no gaps.
Execute: scripts/renumber-tasks.sh
Report the result to the user.
Prime the current session with essential project context.
Read and follow references/prime.md. The context/focus area is everything after "prime" in the user input above (if anything).