# AI use cases for DSH Research Report (PerryLink/dsh-research-report)

Repo: [PerryLink/dsh-research-report](https://github.com/PerryLink/dsh-research-report) | Use cases: 4 | Page: https://dedupedai.com/use-cases/perrylink-dsh-research-report/

dsh-research-report is a DeepSeek Harness plugin that binds each report claim to immutable evidence snapshots, verifies those claims byte-for-byte, and seals the report with a recomputable manifest hash.

This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL.

- Seal a research report with byte-checked claims: https://dedupedai.com/use-cases/perrylink-dsh-research-report/seal-a-research-report-with-byte-checked-claims/
- Verify a sealed report from its directory alone: https://dedupedai.com/use-cases/perrylink-dsh-research-report/verify-a-sealed-report-from-its-directory-alone/
- Find claims in a draft that lack source evidence: https://dedupedai.com/use-cases/perrylink-dsh-research-report/find-claims-that-lack-source-evidence/
- Track claims that the evidence has disproven: https://dedupedai.com/use-cases/perrylink-dsh-research-report/track-claims-that-evidence-has-disproven/

## How to seal a research report with byte-checked claims with DSH Research Report

Repo: [PerryLink/dsh-research-report](https://github.com/PerryLink/dsh-research-report) | Difficulty: Intermediate | Time: 45 minutes | Basis: From the README

Bind each claim to saved source snapshots and seal a report whose manifest hash anyone can recompute.

**Who it is for:** A researcher or analyst who must hand over a report where every number and quoted span traces back to saved sources.

**The problem:** Report numbers and quotes drift away from their sources, and readers cannot tell which ones were checked.

**The outcome:** A sealed report folder with a seal hash, per-claim verdicts, and a manifest that records every evidence hash.

### Steps

1. Run `dsh plugin --profile demo add dsh-research-report` and restart the harness so the plugin takes effect.
2. Register each source with `evidence_add({ origin: "<file or URL>", title: "<source name>" })` and keep the returned evidence id. (inferred, check the README)
3. Call `research_report({ topic: "<topic>", sections: [...], claims: [{ id: "c1", text: "<claim text>", evidenceIds: ["<evidence id>"] }], evidenceRefs: ["<evidence id>"] })` to assemble and seal the report. (inferred, check the README)
4. Check the summary line for counts of verified, unverified, contradicted, insufficient, and disproven claims.
5. Run `ledger_query({ claimId: "c1" })` to see the bindings and the latest verdict for a claim.

### Prompt to paste into your agent

```text
Using the dsh-research-report plugin in DeepSeek Harness, register the source at <source path or URL> with evidence_add, then seal a report on <topic> whose claims bind to that evidence. Expected output: a seal hash and a claim count line. Verify that every claim you wrote shows a verdict and that no claim with an empty evidenceIds list is marked verified.
```

### Caveat

Verification is a byte check: it confirms that numbers and quoted spans appear in the snapshots, not that the claims are semantically correct.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/perrylink-dsh-research-report/seal-a-research-report-with-byte-checked-claims/

## How to verify a sealed report from its directory alone with DSH Research Report

Repo: [PerryLink/dsh-research-report](https://github.com/PerryLink/dsh-research-report) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README

Recompute a report's seal and per-claim checks with the standalone verifier, without the original session.

**Who it is for:** A reviewer or auditor who receives a sealed report folder and must confirm it has not been altered.

**The problem:** A reader has to trust the report on faith unless they can recompute its hashes and re-check its claims independently.

**The outcome:** A JSON or SARIF result showing whether the seal hash and each claim re-check still hold.

### Steps

1. Install the plugin with `dsh plugin --profile demo add dsh-research-report` so the verifier is available.
2. Locate the sealed report directory, which contains report.md, manifest.json, verification.jsonl, and disconfirmation.jsonl.
3. Run `dsh-research-verify --report <report-dir> --format json` to get a JSON envelope with the recomputed results. (inferred, check the README)
4. Run `dsh-research-verify --report <report-dir> --format sarif` if your tooling reads SARIF 2.1.0. (inferred, check the README)
5. Compare the recomputed seal hash with the seal hash printed when the report was sealed.

### Prompt to paste into your agent

```text
Use the dsh-research-verify CLI from the dsh-research-report repo on the sealed report at <report-dir>. Expected output: a JSON envelope or SARIF document that reports the recomputed seal hash and per-claim results. Verify that the recomputed seal hash matches the one recorded in manifest.json before reporting the report as intact.
```

### Caveat

The verifier recomputes hashes and re-runs byte checks from the sealed directory, so it cannot judge whether the sources themselves were right.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/perrylink-dsh-research-report/verify-a-sealed-report-from-its-directory-alone/

## How to find claims in a draft that lack source evidence with DSH Research Report

Repo: [PerryLink/dsh-research-report](https://github.com/PerryLink/dsh-research-report) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README

Seal a draft and read the visible markers to see which claims have no bound evidence.

**Who it is for:** A writer or analyst reviewing an AI-drafted document who needs to know which statements have no backing source.

**The problem:** Unsupported statements look as authoritative as sourced ones, so gaps are easy to miss.

**The outcome:** A sealed report where every unsupported claim carries a visible marker and appears in Appendix A.

### Steps

1. Run `dsh plugin --profile demo add dsh-research-report` if the plugin is not yet installed.
2. Submit the draft through `research_report({ topic: "<topic>", sections: [...], claims: [{ id: "c1", text: "<claim>", evidenceIds: [] }], evidenceRefs: [] })`. (inferred, check the README)
3. Open the generated report.md and search for the unverified marker in the report body.
4. Read the Appendix A list in report.md and note each claim it names.
5. Register sources with `evidence_add` and reseal the report with the new evidence ids bound to those claims.

### Prompt to paste into your agent

```text
Using dsh-research-report in DeepSeek Harness, seal a report on the draft at <draft path> with no evidence bound to the claims. Expected output: report.md with unverified markers on each unsupported claim. Verify that every claim with an empty evidenceIds list is marked unverified rather than passed silently.
```

### Caveat

A claim with no evidence is marked unverified, and the plugin cannot find the missing sources for you.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/perrylink-dsh-research-report/find-claims-that-lack-source-evidence/

## How to track claims that the evidence has disproven with DSH Research Report

Repo: [PerryLink/dsh-research-report](https://github.com/PerryLink/dsh-research-report) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README

Query the ledger and the falsification log to see which claims the sources contradict.

**Who it is for:** An analyst preparing a fact-check who must list every claim the saved evidence contradicts.

**The problem:** Contradicted claims get lost among sourced ones, and rewritten text can repeat a disproven claim unnoticed.

**The outcome:** A list of contradicted and disproven claims with their evidence references and reasons from disconfirmation.jsonl.

### Steps

1. Run `ledger_query({ claimId: "<claim id>" })` to read the bindings and latest verdict for one claim. (inferred, check the README)
2. Run `ledger_query({})` with no id to get a ledger summary.
3. Open disconfirmation.jsonl in the sealed report folder to read each contradicted or disproven claim with its evidence references and reason.
4. Check the falsification log appendix in report.md for the same entries.

### Prompt to paste into your agent

```text
Use ledger_query from dsh-research-report on the claims in <report-dir>. Expected output: a list of contradicted and disproven claims with their reasons. Verify that each listed claim also appears in disconfirmation.jsonl before you report it.
```

### Caveat

Negative knowledge only re-marks a repeated claim as disproven while the evidence is unchanged, so the list reflects the saved snapshots, not new research.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/perrylink-dsh-research-report/track-claims-that-evidence-has-disproven/