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 skills add zebbern/agent-collab --skill cursor-promptinggit clone --depth 1 https://github.com/zebbern/agent-collabWrote 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/skills/zebbern/agent-collab/cursor-prompting)<a href="https://agentmods.dev/skills/zebbern/agent-collab/cursor-prompting"><img src="https://agentmods.dev/badge/skills/zebbern/agent-collab/cursor-prompting/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/zebbern/agent-collab/cursor-prompting"><img src="https://agentmods.dev/badge/skills/zebbern/agent-collab/cursor-prompting.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.00031 | $0.00626 |
| Opus 5.5 | $0.00012 | $0.00250 |
| Sonnet 5.5 | $0.00006 | $0.00125 |
| Haiku 4.5 | $0.00003 | $0.00063 |
Grade A, and why
cursor-prompting 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 1mo 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.
What it actually says
Cursor Prompting
Use this skill when cursor:cursor-rescue needs to shape a request before forwarding it to cursor-agent.
Prompt Cursor like an operator, not a collaborator. Keep prompts compact and block-structured. State the task, the output contract, the follow-through defaults, and the small set of extra constraints that matter. Cursor's default auto router may hand the prompt to any model family (GPT, Claude, Grok, Composer), so avoid model-specific idioms and rely on plain structure that every family parses well.
Prompt shape:
- Lead with one sentence naming the goal and the definition of done.
- List hard constraints (files that must not change, behavior that must be preserved, tests that must pass).
- State the expected output: a diff, a diagnosis with evidence, a ranked list of findings, or a plan.
- For fixes: tell it to run the relevant tests and report the results, not just claim success.
- For diagnosis: ask for observed facts, inferences, and open questions as separate sections.
- Do not paste large file contents into the prompt — Cursor reads the workspace itself.
Model selection (only when the user asks, otherwise leave auto):
--profile deep|fastontask/rescueis a coarse, named shortcut (deep=gpt-5.6-sol-xhigh,fast=cursor-grok-4.5-high-fast) with no effort tier — Cursor has none.--modelstill overrides it and remains the explicit-model escape hatch.auto— Cursor's server-side router; the right default for most tasks.- Deep or adversarial reviews: a thinking model such as
claude-opus-5-thinking-high. - Quick, tightly bounded edits: a fast Codex tier such as
gpt-5.3-codex-fastorgpt-5.3-codex-low-fast. - Long-context work across many files: a 1M-context variant (
-highthinking models,gpt-5.6-sol-high). - The live roster changes;
cursor-agent --list-modelsis the source of truth. Pass names through--modelunchanged.
Mode notes:
- Reviews are requested read-only and the plugin reports any git-visible workspace drift afterward, but Cursor has no enforced read-only sandbox — never promise edits in a review prompt, and never rely on a review being unable to write.
- Write-capable tasks run with
--forceunder the plugin's ownership-tracked cancel — still scope the prompt to the smallest safe change. - Resume (
--resume <chat-id>) continues an existing Cursor session with its context; write follow-up prompts as instructions to continue, not as restatements of the whole task.
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.
- 1mo ago First seen · 33 lines · 31 tokens per session scan A 8317b7461736
cursor-prompting is a skill published in the GitHub repository zebbern/agent-collab (34 stars, last pushed 25d ago), licensed Apache-2.0. It adds 31 tokens to every session and 626 once invoked, about $0.0001 per session on Opus 5.5. 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 skills, from other repositories
llm-app-development
LLM app development with RAG, prompt engineering, vector databases, and AI agents.
prompt-factory
World-class prompt powerhouse that generates production-ready mega-prompts for any role, industry, and task through intelligent 7-question flow, 69 comprehensive presets across 15 professional domains (technical, business, creative, legal, finance, HR, design, customer, executive, manufacturing, R&D, regulatory…
rootnode-block-selection
Guides selection of identity, reasoning, and output approaches for Claude prompts based on task characteristics. Trigger on: "help me choose an approach," "which approach fits this task," "recommend a prompt pattern," "compare reasoning methods," "map this task to the right approach," "what combination of approaches,"…
rootnode-domain-agentic-context
Specialized agentic and context engineering prompt methodology for Claude. Use when designing AI agents — system prompts, tool definitions, context window architectures, failure mode planning, multi-agent coordination, agent evaluation. Trigger on: "design an agent," "build a system prompt for an agent," "architect an…
rootnode-prompt-validation
Evaluates and scores Claude prompts and system prompts using a six-dimension Scorecard with anchored 1-5 rubrics. Trigger on: "review my prompt," "score this prompt," "rate my prompt," "evaluate this prompt," "improve my prompt," "what's wrong with my prompt." Also trigger on: "why isn't my prompt working," "my prompt…
rootnode-domain-content-communications
Specialized prompt-building methodology for content and communications tasks. Provides 11 tested approaches (3 identity, 4 reasoning, 4 output) for writing, editing, content strategy, copywriting, audience analysis, editorial evaluation, cross-channel adaptation, and persuasion design. Use when building a Claude…