concept:research

A web-research command for concept development that crawls supplied pages, keeps content relevant to a search query, and registers the findings as sources. crawl4ai is a tool for collecting and processing content from websites.

In plain words
What is it for?
Use it for interactive research, single-page searches, batches of URLs, or deeper crawls of documentation sites. You can provide a topic and URLs, or begin without them and answer prompts.
Why use it?
It reduces the need to read every part of a web page when only some content relates to your topic. It also records research findings in the concept project's source registry.

Command

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 commands/ddunnock/claude-plugins/concept.research
Clone the repo
git clone --depth 1 https://github.com/ddunnock/claude-plugins
Per session 33 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,241 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 $0.00033 $0.01241
Opus 5 $0.00016 $0.00620
Sonnet 5 $0.00007 $0.00248
Haiku 4.5 $0.00003 $0.00124

Measured 2d ago against content hash 9e910f759287, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

concept:research 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.

skills/concept-dev/commands/concept.research.md · 144 lines

How it starts

The opening of the file, as written. The whole thing — 144 lines — stays where its author put it; the contents beside it link to each section on GitHub.

/concept:research

Research web sources for concept development. Crawls URLs with relevance filtering and automatically registers findings in the source registry.

Usage

/concept:research [url] [query]
/concept:research                         # Interactive — prompts for topic and URLs
/concept:research https://example.com/docs "spacecraft thermal management"

Procedure

Step 1: Gather Research Parameters

If no arguments provided, ask the user:

  1. Research topic/query — What are you researching? (used for BM25 relevance filtering)
  2. URL(s) — Specific URLs, or should we start with a WebSearch?
  3. Mode — Single page, batch, or deep crawl of a documentation site?

If URL(s) and query are provided as arguments, proceed directly.

Step 2: Check crawl4ai Availability

Verify crawl4ai is installed:

python3 -c "import crawl4ai; print(f'crawl4ai {crawl4ai.__version__}')"

If not installed, inform the user:

crawl4ai is not installed. Install it from PyPI with: pip install crawl4ai (Verify the package at https://pypi.org/project/crawl4ai/ before installing.)

In the meantime, I can use WebSearch and WebFetch for research.

Fall back to WebSearch + WebFetch workflow if crawl4ai is unavailable.

Step 3: Determine Current Phase

Check state.json for the current phase:

python3 ${CLAUDE_PLUGIN_ROOT}/scripts/update_state.py --state .concept-dev/state.json show

Use the current phase for source tagging. Default to drilldown if no session is active.

Step 4: Execute Research

Choose the appropriate mode based on user input:

Single URL — focused deep-read:

python3 ${CLAUDE_PLUGIN_ROOT}/scripts/web_researcher.py crawl "<url>" --query "<query>" --phase <phase>

Use --css-selector if the user wants to focus on a specific page section (e.g., .main-content, article).

Multiple URLs — batch crawl:

Read the full file on GitHub · 144 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 · 144 lines · 33 tokens per session scan A 9e910f759287

Subscribe to this mod's changes

concept:research is a command published in the GitHub repository ddunnock/claude-plugins (12 stars, last pushed 5mo ago), licensed MIT. It adds 33 tokens to every session and 1,241 once invoked, about $0.0002 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.