Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add rules/wanghao9610/star/skill-rootsgit clone --depth 1 https://github.com/wanghao9610/STARWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/rules/wanghao9610/star/skill-roots)<a href="https://agentmods.dev/rules/wanghao9610/star/skill-roots"><img src="https://agentmods.dev/badge/rules/wanghao9610/star/skill-roots.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00744 | $0.00744 |
| Opus 5 | $0.00372 | $0.00372 |
| Sonnet 5 | $0.00149 | $0.00149 |
| Haiku 4.5 | $0.00074 | $0.00074 |
Grade A, and why
skill-roots scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 44 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Guidelines for loading skills from tool-specific roots without mixing incompatible instructions.
Tradeoff: These guidelines favor predictable tool behavior over treating every discovered skill as interchangeable.
1. Respect Root Ownership
Use one authoritative copy per tool. Keep tool-specific instructions separate.
Cursor scans .cursor/skills/ and .agents/skills/ as native roots, plus .claude/skills/ and .codex/skills/ for compatibility. It surfaces one entry per skill name, so when the same skill exists in several roots only one copy is presented to the agent, and it is not necessarily the Cursor-owned one — in this repo all fifteen star-* skills are surfaced from .agents/skills/, the tool-neutral copies. Section 2 is what corrects that. This repo gives each tool-specific root an owner, and the shared root none:
| Root | Owner tool | Adaptation |
|---|---|---|
.cursor/skills/ |
Cursor | SwitchMode plan gate, Task subagents, AskQuestion |
.agents/skills/ |
none — the shared AGENTS.md root every agent reads |
no tool named at all: bare star-* invoke, "your plan tool", "a read-only sub-agent" / "a writing sub-agent", "your question tool" |
.claude/skills/ |
Claude | /skill invoke, AskUserQuestion, Enter/ExitPlanMode |
Do not treat these three trees as interchangeable. Keep tool-specific wording in the matching root only.
2. Follow the Cursor-Owned Copy
When duplicate skills exist, Cursor Agent must use the copy under .cursor/skills/.
- For
star-plan-*, and any other skill that exists in multiple roots, read and follow.cursor/skills/<name>/. - This holds even when the surfaced entry points somewhere else: the path shown in the skill listing is where Cursor found a copy, not the copy to act on. Open
.cursor/skills/<name>/SKILL.mdyourself and follow that. The surfaced copy names no Cursor tool — "your question tool" where.cursorsaysAskQuestion— so acting on it costs you the Cursor-specific instructions. - Ignore sibling copies under
.agents/skills/and.claude/skills/unless the user explicitly asks for the tool-neutral or the Claude wording. - When linking hand-offs, keep
/star-plan-*names, but still load instructions from.cursor/skills/.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 3d ago First seen · 44 lines · 744 tokens per session scan A 01dd5e6ba049
skill-roots is a cursor rule published in the GitHub repository wanghao9610/STAR (51 stars, last pushed 2d ago), licensed MIT. It adds 744 tokens to every session, about $0.0037 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other cursor rules, from other repositories
coding-standards
Default comments in English; Python prefers single quotes.
development-conventions
Development conventions, code style, and best practices for the Kaneo project.
checkpoint
Write a committed handoff checkpoint so a fresh session can resume with zero loss. Use when crossing the 40% context line, on every stage transition, and at session end. Writes .trace/checkpoints/ - .md while the agent is still sharp.
conservative-file-creation
Check for existing files before creating. Be conservative about new files and code.
repository-layout
Canonical map of repo folders and docs.
angular-20
This rule provides comprehensive best practices and coding standards for Angular development, focusing on modern TypeScript, standalone components, signals, and performance optimizations.