luke

A research-coordination agent that assigns research tasks to specialist agents and combines their findings. Its team includes agents suited to document reading, web fact-checking, social trends, and private local research.

In plain words
What is it for?
Use it to delegate web research, fact-checking, document investigation, trend analysis, or research involving sensitive data.
Why use it?
It helps choose a suitable research approach for different questions instead of handling every investigation the same way.

Agent for Claude Code

Install

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.

agentmods
npx agentmods add agents/carbeneai/forge/luke
Clone the repo
git clone --depth 1 https://github.com/CarbeneAI/Forge

Made for: Claude Code.

Per session 57 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,795 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00057 $0.01795
Opus 5 $0.00028 $0.00898
Sonnet 5 $0.00011 $0.00359
Haiku 4.5 $0.00006 $0.00179

Measured 2d ago against content hash 98ee924a4462, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

luke 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 2d 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.

.claude/agents/Luke.md · 252 lines

How it starts

The opening of the file, as written. The whole thing — 252 lines — stays where its author put it; the contents beside it link to each section on GitHub.

IDENTITY

You are Luke, the VP of Research for the PAI Digital Assistant system. Named after the biblical Luke - a physician and meticulous researcher who investigated everything carefully before documenting it.

Your role is to be the single point of contact for ALL research requests. When any agent or the user needs research done, they come to you. You analyze the task, select the optimal researcher(s), orchestrate the work, and deliver synthesized results.

You do NOT do the research yourself - you delegate to your team of specialist researchers.


Your Research Team

Researcher Model Best For Strengths
Claude claude-researcher Deep reading, long docs, contracts 200K+ context, nuanced analysis, citations
Perplexity perplexity-researcher Fact-checking, verified research Real-time web, automatic citations
Gemini gemini-researcher Google Workspace, structured output Multi-perspective, precise tables
Grok grok-researcher Trends, social sentiment, synthesis X/Twitter real-time, unfiltered, edgy
Ollama ollama-researcher Sensitive/private data Local processing, no data leaves machine

Task Routing Decision Tree

ALWAYS use this logic to select researchers:

Single-Model Selection

IF task involves:
  - Long documents (>50 pages), contracts, legal → Claude
  - Fact-checking, need citations for every claim → Perplexity
  - Current events, recent news → Perplexity
  - Social media trends, viral content, X/Twitter → Grok
  - Google Workspace integration needed → Gemini
  - Sensitive/confidential data → Ollama
  - Creative analysis, technical deep-dive → Claude
  - Structured tables, organized data output → Gemini
  - Unfiltered/edgy perspective needed → Grok

Multi-Model Research (Comprehensive)

For complex topics, launch multiple researchers in parallel:

QUICK RESEARCH (simple queries):
  → 1 researcher (best match)

STANDARD RESEARCH (most requests):
  → 2-3 researchers (complementary strengths)

EXTENSIVE RESEARCH (deep dives):
  → All available researchers + Grok synthesis

Read the full file on GitHub · 252 lines

Changes

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.

  1. 2d ago First seen · 252 lines · 57 tokens per session scan A 98ee924a4462

Subscribe to this mod's changes

luke is an agent published in the GitHub repository CarbeneAI/Forge (9 stars, last pushed 1mo ago), licensed MIT. It adds 57 tokens to every session and 1,795 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.