coding
Manage coding sessions. Use when the OPERATOR asks to work on code, hack on a repo, implement features, fix bugs, or do any development work in a codebase.
- License 许可
- MIT
- 所属仓库
- mrtolkien/GHOST
- 最近更新
- 2026-04-24
- Stars
- ★ 5
- Forks
- 0
01 前言 Front Matter
- name
- coding
- description
- Manage coding sessions. Use when the OPERATOR asks to work on code, hack on a repo, implement features, fix bugs, or do any development work in a codebase.
03 SKILL.md 正文
Coding Session Manager
Use this skill when the OPERATOR asks for coding work on a repository.
What This Does
You spawn a coding agent that takes over the Discord channel. The coding agent has
full access to the repo, can read/edit files, run commands, and commit changes. The
OPERATOR talks directly to the coding agent until they send /kill to end the session.
Don't create a project for single-session coding tasks. Only suggest a project if the work will span multiple sessions or has 3+ distinct tasks to track.
Workflow
1. Identify the Repo
Check if the OPERATOR has a project with a repo.md reference note:
knowledge_search query="repo.md <project-name>" scope="references"
If found, the repo.md will contain the repo URL and local path.
2. Set Up the Code Directory
Repos live under code/ in the workspace. Use the project slug as the directory name.
If not cloned yet:
shell command="git clone <repo-url> code/<slug>"
If already cloned:
shell command="cd code/<slug> && git fetch && git status"
3. Start the Coding Session
shell command="ghost hack start code/<slug> --prompt '<task description>'"
The command outputs:
coding_session_id=<id>
session_id=<id>
working_dir=<path>
4. Explain to the OPERATOR
After starting, tell the OPERATOR:
"I've started a coding session on
<repo>. You're now talking to the coding agent — it has full access to the repo. Send/killwhen you're done."
5. After /kill
When the OPERATOR ends the session, you'll receive a system message with a git summary (branch, commits, changed files). Use this to:
- Update the project's task status if applicable
- Write a diary entry about what was accomplished
- Note any follow-up items the OPERATOR mentioned
6. Resume a Previous Session
If the OPERATOR wants to continue previous work:
shell command="ghost hack list"
Then resume with:
shell command="ghost hack resume <coding_session_id>"
Creating repo.md References
When setting up a new repo for the first time, create a reference note:
note_write title="repo.md" project="<slug>" content="..."
Include: repo URL, local path (code/<slug>), branch conventions, build/test commands,
and any setup notes.
Important
- Always check for an existing
repo.mdbefore cloning - The channel ID is passed automatically — no need to specify it manually
- The coding agent has its own session — your conversation history is separate
- After
/kill, the coding agent's session is preserved for futureresume
04 同仓库更多 Skills
system-management
Use when reading or changing GHOST configuration (providers, models, timing, compaction, services, any config.toml key), adding or switching LLM providers, answering questions about GHOST's own features or capabilities, a CLI tool is missing or "not found", services need starting/stopping/debugging, updating GHOST to a newer version, managing the Nix shell environment, or troubleshooting GHOST setup.
writing-skills
Use when creating new skills, editing existing skills, or verifying skills work before deployment
writing-plans
Use when you have a spec or requirements for a multi-step task, before touching code. Write comprehensive implementation plans assuming the engineer has zero context for our codebase and questionable taste.
verification
Use when about to claim work is complete, fixed, or passing, before committing or creating PRs. Requires running verification commands and confirming output before making any success claims. Evidence before assertions, always.
tdd
Use when implementing any feature or bugfix, before writing implementation code. Write the test first. Watch it fail. Write minimal code to pass.
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes