customize
Create a Wisp specialist or author a project-local skill using the tools Wisp actually exposes. Use when the user wants a custom Agent persona, a restricted specialist loadout, a new skill, or changes to an existing project-local skill.
Install
npx skills add https://github.com/xuzhougeng/wisp-science/tree/main/skills/customize
claude plugin marketplace add https://llmmart.ai/marketplace.json && claude plugin install xuzhougeng-wisp-science@llmmart
git clone https://github.com/xuzhougeng/wisp-science.git
The skills CLI installs just this skill, for any of its supported agents. Claude Code installs the whole xuzhougeng/wisp-science collection as a plugin from our marketplace. Git is the plain clone.
Skill manifest
Customize Wisp
Wisp exposes customization through explicit tools and Settings UI. Python has no control-plane SDK. Never invent profile, connector, or skill CRUD methods.
Create a specialist
Use save_specialist only when it is advertised in the current session.
- Ask what single job the specialist owns and what it must not do.
- Draft a display name, one-line description, and concise instructions that extend Wisp's base prompt.
- Decide whether to inherit project skills/connectors or use explicit
whitelists. Use exact installed skill names returned by
search_skills. Include connector names only when the user supplied or confirmed them. - Show the complete proposal and obtain confirmation before creating or updating it.
- To edit an existing specialist, call
configurewithaction=getand keyspecialists, thensave_specialistwith thatid. Omitidto create. Builtin instruction text cannot be replaced. - Call
save_specialistwithnameandinstructions; adddescription,model_id,skills, orconnectorsonly when intentionally chosen. - Report the returned specialist id. Deletion, icons, and color still live under Settings → Specialists.
Omit skills and connectors to inherit project settings. An empty list is an
explicit zero-access whitelist and is not equivalent to omission.
Author or update a skill
Load skill-creator. For a project-local skill, edit files under:
.wisp/skills/<skill-name>/SKILL.md
Use read, write, and edit. search_skills and use_skill are discovery
and loading tools, not mutation tools. If a newly created project skill is not
visible immediately, refresh or reopen the project so Wisp rebuilds its skill
index.
For a user-wide skill, create the folder in the project first, validate it, then ask the user to install that folder through Settings → Skills. Wisp copies installed skills to its user skill directory; there is no Agent-side publish or delete interface.
Custom appearance
For a user stylesheet (hide the lead bar, restyle chat, override tokens), load
custom-theme. Write CSS that uses documented tokens, then apply it with
configure set custom_css or ask the user to import the file under
Settings → Appearance.
Boundaries
- Do not modify Wisp's SQLite store directly.
- Do not use
python,shell, orrun_in_contextto bypass Settings or tool authorization. - Do not promise conversation identity switching; the user selects specialists through Wisp's UI.
- Do not claim connector enumeration when no connector-management tool is advertised.
Files (wisp-science)
-
SKILL.md 2.9 KB
--- name: customize description: Create a Wisp specialist or author a project-local skill using the tools Wisp actually exposes. Use when the user wants a custom Agent persona, a restricted specialist loadout, a new skill, or changes to an existing project-local skill. --- # Customize Wisp Wisp exposes customization through explicit tools and Settings UI. Python has no control-plane SDK. Never invent profile, connector, or skill CRUD methods. ## Create a specialist Use `save_specialist` only when it is advertised in the current session. 1. Ask what single job the specialist owns and what it must not do. 2. Draft a display name, one-line description, and concise instructions that extend Wisp's base prompt. 3. Decide whether to inherit project skills/connectors or use explicit whitelists. Use exact installed skill names returned by `search_skills`. Include connector names only when the user supplied or confirmed them. 4. Show the complete proposal and obtain confirmation before creating or updating it. 5. To edit an existing specialist, call `configure` with `action=get` and key `specialists`, then `save_specialist` with that `id`. Omit `id` to create. Builtin instruction text cannot be replaced. 6. Call `save_specialist` with `name` and `instructions`; add `description`, `model_id`, `skills`, or `connectors` only when intentionally chosen. 7. Report the returned specialist id. Deletion, icons, and color still live under **Settings → Specialists**. Omit `skills` and `connectors` to inherit project settings. An empty list is an explicit zero-access whitelist and is not equivalent to omission. ## Author or update a skill Load `skill-creator`. For a project-local skill, edit files under: ```text .wisp/skills/<skill-name>/SKILL.md ``` Use `read`, `write`, and `edit`. `search_skills` and `use_skill` are discovery and loading tools, not mutation tools. If a newly created project skill is not visible immediately, refresh or reopen the project so Wisp rebuilds its skill index. For a user-wide skill, create the folder in the project first, validate it, then ask the user to install that folder through **Settings → Skills**. Wisp copies installed skills to its user skill directory; there is no Agent-side publish or delete interface. ## Custom appearance For a user stylesheet (hide the lead bar, restyle chat, override tokens), load `custom-theme`. Write CSS that uses documented tokens, then apply it with `configure` `set custom_css` or ask the user to import the file under **Settings → Appearance**. ## Boundaries - Do not modify Wisp's SQLite store directly. - Do not use `python`, `shell`, or `run_in_context` to bypass Settings or tool authorization. - Do not promise conversation identity switching; the user selects specialists through Wisp's UI. - Do not claim connector enumeration when no connector-management tool is advertised.
Comments (0)
Sign in to join the conversation.
Reviews (0)
No reviews yet.
No comments yet.