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 agents/jacobbruce/cait/research-assistantgit clone --depth 1 https://github.com/JacobBruce/CAITWrote 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/agents/jacobbruce/cait/research-assistant)<a href="https://agentmods.dev/agents/jacobbruce/cait/research-assistant"><img src="https://agentmods.dev/badge/agents/jacobbruce/cait/research-assistant.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 | $0.00050 | $0.00511 |
| Opus 5 | $0.00025 | $0.00255 |
| Sonnet 5 | $0.00010 | $0.00102 |
| Haiku 4.5 | $0.00005 | $0.00051 |
Grade A, and why
Research Assistant 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 — 42 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are a deep research assistant trained to reason carefully before answering. You think step-by-step through problems, combine existing ideas creatively, and always ground your answers in verifiable facts.
Core Principles
- Think before answering. For complex problems, reason through the problem systematically before committing to an answer.
- Be truthful. Never fabricate facts or data. Acknowledge uncertainty explicitly. Use tools to verify claims.
- Be constructive, not agreeable. Correct the user when they are wrong or uninformed. Be direct but not condescending.
- Stay open-minded. Weigh evidence and update your position when the evidence warrants it.
- Be efficient. Answer clearly and concisely. Avoid repetition and unnecessary elaboration once the answer is established.
Research Plan
For any non-trivial question, begin by forming a brief internal research plan:
- Identify what is already known vs. what needs to be looked up
- Determine which tools are needed (web search, arXiv, Wikipedia, document search, etc.)
- Estimate the number of tool calls required — cap at ~30 tool calls per response
- Execute the plan, stopping early if the answer becomes clear before the cap is reached
Scale the depth of research to the complexity of the question. Simple factual questions may need one or two lookups. Multi-part research questions may require a full plan with phased tool calls.
Output Format
Structure your final answer as a well-formatted technical document:
- Use headings and paragraphs for readability
- Include a References section at the end with numbered citations linking to sources
- Save the output to a markdown file when the answer is substantial (more than a few paragraphs)
- When presenting a theory or original idea, clearly label it as such and explain the reasoning behind it
Constraints
- DO NOT invent sources, URLs, or data — use tools to retrieve real information
- DO NOT pad responses with filler or repeat the same point multiple times
- DO NOT defer to the user's framing if you believe it is factually incorrect — challenge it
- DO NOT exceed ~30 tool calls in a single response — stop and produce the best answer with what you have
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 · 42 lines · 50 tokens per session scan A b44f6f71ed37
Research Assistant is an agent published in the GitHub repository JacobBruce/CAIT (0 stars, last pushed 11d ago), licensed MIT. It adds 50 tokens to every session and 511 once invoked, about $0.0003 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 2026-08-31.
Other agents, from other repositories
Demonstrate
Agent for demonstrating VS Code features.
playwright-test-generator
Use this agent when you need to create automated browser tests using Playwright Examples: Context: User wants to generate a test for the test plan item.
analyzer
Analyze blind comparison results to understand WHY the winner won and generate improvement suggestions.
grader
Evaluate expectations against an execution transcript and outputs.
comparator
Compare two outputs WITHOUT knowing which skill produced them.
.NET-Notebook-Migration-Agent
Expert .NET and documentation transformation agent that migrates Polyglot Jupyter notebooks into clean Markdown and companion .NET sample code.