- Home
- Skills
- Research & Analysis
- Research Claim Verification Pass
More screenshots
Works with the AI Agents you already use
Research Claim Verification Pass
Check a draft claim by claim, tag each by evidence status, and keep measured results separate from inference.
$5
Research Claim Verification Pass
Example session with this skill installed
Apply the skill to: Briefing draft with a hidden weak claim.
- Read your context and instructions
- Compiled the research claim verification
Worked example: A briefing draft argued for a programme expansion, citing a 34 percent improvement, a cost comparison, and a causal claim about retention. Findings: The 34 percent figure traced to a summary slide, not to primary data, and the primary table showed 21 percent on the comparable measure. The cost comparison mixed a pilot cost with a full-scale cost, overstating the saving. The retention claim used causal language from a correlational survey, so the verb was softened to associated. Two scope claims used all participants where the sample covered one site only. Outcome: The corrected draft reported 21 percent with its table location, a like-for-like cost comparison, and associated language for retention. The expansion recommendation survived on the corrected numbers, but with a narrower scope and a note that a second site was needed before generalising. This is a worked example produced by applying the skill protocol, not a guarantee of results on new data.
Connects securely to your tools. The creator never sees your data.
About this skill
Drafts mix measured results, cited findings, and author interpretation in the same sentences, so readers cannot tell which parts would survive if one source were withdrawn. The bottleneck is claim extraction. Writers verify whole paragraphs, which lets a single weak claim hide inside a strong paragraph. Verification must happen claim by claim, with each claim tagged by evidence status. This skill gives you a full protocol: trigger and scope, an inputs and outputs contract, ordered steps with checks, a decision table, a fill-in template, a checklist, a scoring rubric, a worked example, edge cases, and a pre-delivery quality gate. It also includes a runnable script under scripts/ so the core scoring or checking step can be repeated on new data without redoing the judgement by hand. The focus throughout is bottleneck fixes and workflows that a short free checklist would not cover: what to check first and second, which thresholds change the action, and when to stop and narrow the scope.
Verification checks what sources say and whether numbers trace; it does not rerun experiments or guarantee a source is correct. A claim can trace cleanly to a flawed source and still be wrong. Results depend on input quality: missing inputs, small samples, or unverified sources limit how strongly conclusions can be stated. The skill helps you state those limits honestly rather than hiding them.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 23 days ago