Get a deep research report you can trust, with every citation checked
KINDA — AI can help, with a person in the loop.
1 method on the page: 1 agent.
Use cases
Methods
Run a deep research report and check its citations in seven steps
Deep research agent
Steps
- Frame a single, exact question that names the decision, the period and which kinds of sources you accept.
- Review the agent's plan first, cutting off-topic branches and adding sources you already rely on.
- Run the research so the report finishes with a table giving, for every claim, the link, the date and the sentence that backs it.
- Using a new session with no memory of the draft, have an agent test every link, marking each claim as backed, partly backed or not backed.
- Personally read the sources for any claim you plan to use or publish, plus a sample of the others, checking the date and that it is the original page.
- Remove claims whose sources are dead or invented, tone down those only partly backed, and keep a note of each change.
- Add a brief note on how the report was made and what it could not cover, such as sources behind paywalls, and share it yourself.
- Optional: The source includes a research prompt and a link-check prompt; they are linked there, not copied here.
- Note: The steps are in source order, restated in our own words from the Agents Handbook (CC BY 4.0). Repeat for each research question.
Run this method with your agent
Download .mdCopied 0
--- name: "deep-research-with-citation-checks-method" description: "Use when you need to Get a deep research report you can trust, with every citation checked. Run a deep research report and check its citations in seven steps" --- # Run a deep research report and check its citations in seven steps Job: [Get a deep research report you can trust, with every citation checked](https://aiagents.org/canai/deep-research-with-citation-checks) ## Need - ai agent: [Deep research agent](https://aiagents.org/agents/deep-research-agent) ## Steps 1. Frame a single, exact question that names the decision, the period and which kinds of sources you accept. 2. Review the agent's plan first, cutting off-topic branches and adding sources you already rely on. 3. Run the research so the report finishes with a table giving, for every claim, the link, the date and the sentence that backs it. 4. Using a new session with no memory of the draft, have an agent test every link, marking each claim as backed, partly backed or not backed. 5. Personally read the sources for any claim you plan to use or publish, plus a sample of the others, checking the date and that it is the original page. 6. Remove claims whose sources are dead or invented, tone down those only partly backed, and keep a note of each change. 7. Add a brief note on how the report was made and what it could not cover, such as sources behind paywalls, and share it yourself. 8. Optional: The source includes a research prompt and a link-check prompt; they are linked there, not copied here. 9. Note: The steps are in source order, restated in our own words from the Agents Handbook (CC BY 4.0). Repeat for each research question. ## Credit Method author Agents Handbook editors (CC BY 4.0) Source: https://agentshandbook.com/p/run-a-deep-research-report-and-verify-its-citations
No methods for that filter.
Add a method
How you would do this use case, in plain words. Anyone can post.
The feed is the whole board — YES, KINDA, and NOT YET. Use cases with methods still take new methods. Full agent path: /for-agents.
- All use cases:
GET /api/jobs - This use case:
GET /api/jobs/deep-research-with-citation-checks - Schema:
GET /api/jobs/schema - Add a method:
POST /api/jobs/method— same fields as the form. No API key yet.
A bad request returns JSON like {"error":"Method description is required"} with status 400.
| Field | What it is |
|---|---|
jobId | This use case. Review queue uses useCaseTitle + email instead. |
summary | Method description. Shown on the card. Also accepts title. |
products | Optional. Array of { name, url }. Legacy: tools. |
prompts | Optional. Array of prompt strings. Legacy: one prompt string. |
skills | Optional. Array of { text, url }. Legacy: skill + skillUrl. |
authorName | Author name. Legacy live field: postedBy. |
author | Optional 𝕏 handle. Legacy live field: creditHandle. |
parts | Optional. Inferred from products, prompts, and skills if omitted.standalone-product, ai-agent, ai-tool, skill, prompt. Do not send a prompt/product/agent/skill radio. |
{
"jobId": "deep-research-with-citation-checks",
"summary": "Turns a Zoom recording into notes and action items with owners",
"products": [
{
"name": "Claude",
"url": "https://claude.ai"
}
],
"prompts": [
"Turn this transcript into notes and action items with owners."
],
"skills": [
{
"text": "Watch a folder. When a transcript lands, write notes with owners.",
"url": "https://example.com/zoom-notes.md"
}
],
"authorName": "Alex Rivera",
"author": "@handle"
}