Bootstrap Cursor skills
Add a Cursor Agent Skill so agents load specialized instructions only when the task matches—not on every chat.
- Framing (taxonomy)
-
Vibe coding bootstrap (this is not a Business Bootstrap child page).
- Install ladder for marketplace MCP + skills plugins
- Catalog
- In-repo shop (agent-rules; avoids duplicating this page)
Choose an on-ramp
A. Configure it yourself
Use when you want exact control or the agent cannot write under ~/.cursor/.
-
Create
~/.cursor/skills/<skill-name>/(never under~/.cursor/skills-cursor/—that tree is Cursor-managed). -
Add
SKILL.mdwith YAML frontmatter (name,description) and a short body. -
Put long directives in sibling files (
news.md,blog.md, …) and link them fromSKILL.md(progressive disclosure). -
Optionally keep a canonical copy in a git repo and junction/symlink into
~/.cursor/skills/<skill-name>/. -
Add at most a one-line pointer in User Rules /
RULES.mdif agents must discover the skill by name; do not paste the full curriculum into always-on rules. -
Start a new agent chat and confirm the skill appears / can be read.
B. Drop the recipe into a permissive agent
Use when the studio already allows the agent to edit the code hive and Cursor config.
Paste (or @-mention) something like:
Install a personal Cursor skill from these directives. Keep the full text out of always-on User Rules. Canonical path: <repo>/skills/<skill-name>/ Junction or copy to ~/.cursor/skills/<skill-name>/ Add only a thin pointer in RULES.md / documentation.md. Follow create-skill guidance: description with WHAT + WHEN; progressive disclosure.
Then attach the source notes (export, gist, or chat). The agent should:
-
Author
SKILL.md+ reference files -
Install into
~/.cursor/skills/ -
Update thin pointers only
-
Avoid dumping the body into the always-on User Rule paste
Same end state as on-ramp A.
Worked example: writing-news-vs-blog
Recorded from a session that turned a Gemini news-vs-blog style export into a durable skill next to agent-rules.
Goal
-
News body copy: inverted pyramid, neutral attribution, no editorializing
-
Blog body copy: thesis early, conversational authority, anti-slop hygiene
-
Keep house title stance (news outward / blog inward) in thin rules
-
Keep body directives in a skill so they stay out of always-on context
Steps taken
-
Read the export and Cursor’s create-skill guidance (
~/.cursor/skills-cursor/create-skill/SKILL.mdis built-in reference; do not author there). -
Chose personal skill + versioned canonical copy, not a full paste into User Rules.
-
Created canonical tree under the personal agent-rules fork::
skills/writing-news-vs-blog/ SKILL.md news.md blog.md published-sources.md
-
Wrote
SKILL.mdas a mode router (news vs blog) with house stance and links tonews.md/blog.md. -
Incorporated published suite pointers (claude-blog, marketingskills, Kimi anti-slop, news researcher/writer) into
published-sources.mdas optional reference—local directives win unless asked otherwise. -
Junctioned for Cursor discovery (Windows)::
cmd /c mklink /J "$env:USERPROFILE\.cursor\skills\writing-news-vs-blog" ` "C:\code\github.com\AMDphreak\.forks\agent-rules\skills\writing-news-vs-blog" -
Added thin pointers only:
-
general/documentation.md— body-copy section pointing at the skill -
general/global.md— optional skill note (like talk-normal) -
RULES.md/ Cursor User Rule — one AI-ops line: skill path; do not paste full directives -
MAIN.md— note thatskills/is not part of MAIN context assembly
-
-
Left existing title guidance (news outward / blog inward) in Docs rules unchanged aside from the pointer.
Skills vs always-on rules (checklist)
-
Curriculum lives under
skills/, not in the User Rule paste -
descriptionstates WHAT and WHEN (third person) -
Long material is one level deep from
SKILL.md -
Titles / constants that must always apply stay in thin rules
-
Published upstream suites are cited as optional, not silently overriding house STYLE
See also
-
Rules Manager — compose/watch for rules packs
-
Composio MCP setup — marketplace plugin that ships
skills/beside MCP