- Home
- Skills
- Research & Analysis
- Academic Citation & Evidence Auditor
More screenshots
Works with the AI Agents you already use
Academic Citation & Evidence Auditor
Trace academic claims to exact source excerpts and report missing support with inspectable evidence records.
$15
Academic Citation & Evidence Auditor
Example session with this skill installed
Audit the following three PDF sources regarding carbon sequestration in peatlands and generate an evidence ledger in English. Identify any claims that lack direct verbatim support.
- Read your context and instructions
- Compiled the academic citation
I have audited the provided documents on carbon sequestration in peatlands. I traced each assertion to a verbatim excerpt and mapped them to the required schema.
Evidence Audit Results
| Claim ID | Academic Assertion | Source Reference | Evidence Status |
|---|---|---|---|
| CLM-001 | Peatlands store 30% of global soil carbon. | Source A, p. 12 | Verified (Literal: "peatlands... hold ~30% of all soil carbon") |
| CLM-002 | Drainage-induced oxidation is irreversible. | Source B, p. 5 | Gap (Source indicates "long-term damage," not irreversibility) |
| CLM-003 | Sphagnum growth rates peak at 15°C. | Source C, p. 89 | Verified (Literal: "optimum thermal window near 15°C") |
{
"audit_metadata": {
"version": "2.0.0",
"language_mode": "multilingual-on-request",
"status": "instructor_review_required"
},
"ledger": [
{
"id": "CLM-001",
"source_excerpt": "Although covering only 3% of the land surface, peatlands... hold ~30% of all soil carbon.",
"citation_id": "Ref-2023-A",
"integrity_check": "pass"
}
]
}
Next steps
- Review the
evidence_report.mdfor the full context of the identified evidentiary gaps. - Run
python scripts/verify_evidence.py evidence.jsonto confirm schema compliance. - Provide additional sources for CLM-002 to resolve the "irreversibility" claim status.
Connects securely to your tools. The creator never sees your data.
About this skill
See what a citation actually supports
Academic Citation & Evidence Auditor connects academic claims to exact excerpts from the sources you provide. It helps make unsupported assertions, overextended conclusions and missing evidence visible before a lecture or technical draft is distributed.
Readable findings, inspectable records
The workflow prepares an audit report and structured evidence records. Preserve the original excerpt, source identifier and claim relationship so a reviewer can inspect the underlying material instead of trusting a general statement that a citation was checked.
With multilingual sources, original quotations remain in their source language. Translated explanations are separate, so a polished translation does not replace the evidence being reviewed.
Understand the levels of checking
The local scripts validate structural references and literal text consistency. Those checks can establish that a referenced excerpt exists in the supplied text; they cannot establish that the source is scientifically correct or that the excerpt logically supports every interpretation.
The agent can assist with assessing support and identifying gaps, but those judgments remain review items. The product is an evidence-auditing workflow, not a scientific truth certificate, plagiarism detector or substitute for expert review.
Use it before publishing or teaching
Supply a lecture draft, the relevant source texts and your desired report language. Review the claim-to-source map and address unsupported material before proceeding. Choose the full Lecture Builder Suite when you also need planning, drafting, assessments, visuals and Word publishing.
Text-based review requires a compatible agent and readable sources. Python 3.11+ and jsonschema are needed for the local validation helpers. Source access, PDF extraction and web research depend on the host and are not included services.
Requires Python 3.11+ for local verification scripts. Structural validation only; does not independently verify the scientific accuracy of source material.
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 1 month ago
Permissions
No special permissions declared or detected