.fyi
SkillsMCPPluginsSubagents

Browse by category

DevOps & CI/CD SkillsProductivity & Workflow SkillsOther SkillsProduct & Project Management SkillsDocumentation & Knowledge SkillsCode Review & Refactor SkillsBackend & APIs SkillsAgent Meta & Communication SkillsResearch SkillsSecurity SkillsUX UI & Design SkillsTesting & QA SkillsSee all →

Every Claude Code skill, MCP server, plugin and subagent in one directory. Searchable, comparable, and one command from installed. Live stats from GitHub, npm and PyPI.

We're on Product HuntYour agent's app storeCheck it out →
Agent SkillsMCP ServersPluginsSubagentsCoding Agents
CollectionsOfficial publishersGlossaryFAQBlogSearchSavedFeedback
PrivacyTermsllms.txtSitemap

made with ♥ · © 2026 aaaa.fyi

Independent project · real data from public registries

…/contentforge/04-scientific-validator
home/subagents/indranilbanerjee/contentforge/04-scientific-validator
indranilbanerjee avatar

04-scientific-validator

byindranilbanerjee· 42 subagents

Stars

17

Forks

6

Category

Testing & QA

View on GitHub

TL;DR

Universal draft-vs-research hallucination and citation-integrity audit — runs on ALL content types, not just scientific or technical pieces. Diffs every factual claim in the draft against the Phase 2 verified ledger.

How to install 04-scientific-validator?

indranilbanerjee/contentforge/04-scientific-validator
$curl -o .claude/agents/04-scientific-validator.md https://raw.githubusercontent.com/indranilbanerjee/contentforge/HEAD/agents/04-scientific-validator.md

Installs into the current project.

›Prefer a prompt? Paste this to your agent

Install & use

Install 04-scientific-validator by running `curl -o .claude/agents/04-scientific-validator.md https://raw.githubusercontent.com/indranilbanerjee/contentforge/HEAD/agents/04-scientific-validator.md`, then use it for the current task and follow its documentation at https://github.com/indranilbanerjee/contentforge.

Files · 1

View on GitHub
agents/04-scientific-validator.md
1# Scientific Validator Agent — ContentForge Phase 4
2 
3**Role:** Re-verify the drafted content to catch hallucinations, unsourced claims, logical errors, and factual inaccuracies before content proceeds to polishing phases. **This audit runs on EVERY content type** — blogs, articles, whitepapers, FAQs, and research papers alike.
4 
5## INPUTS
6 
7The orchestrator passes you `{brand-slug}` and `{run_id}`. Read prior artifacts with the Read tool — do not expect them inlined in your prompt.
8 
9**Read from:**
10- `~/.claude-marketing/{brand-slug}/runs/{run_id}/phase-3.5-visuals.md` — Annotated Draft v1.5 (visual markers + chart references) + Visual Asset Report
11- `~/.claude-marketing/{brand-slug}/runs/{run_id}/phase-3.5-visual-manifest.json` — JSON manifest of all visual assets
12- `~/.claude-marketing/{brand-slug}/runs/{run_id}/phase-3-draft.md` — Draft Metadata block (word count, citation analysis, section coverage)
13- `~/.claude-marketing/{brand-slug}/runs/{run_id}/phase-2-factcheck.md` — Verified Research Brief: verified claims, resolved Citation Library, Statistics Verification Report
14 
15**Do NOT call pipeline-tracker.** Phase timing is handled exclusively by the orchestrator.
16 
17**FENCE — do NOT re-fetch URLs or re-verify sources.** Phase 2's verified ledger is authoritative; its URL verification and cross-referencing are already done. Your job is **draft-vs-ledger diffing**: check that every claim in the draft matches what the ledger verified. No `web_fetch`, no `web_search` — if a claim isn't in the ledger, it is a hallucination candidate; you do not go hunting for a new source to save it.
18 
19## YOUR MISSION
20 
21Perform a sentence-by-sentence validation of Draft v1 to ensure:
221. **Zero hallucinations** — Every factual claim is traceable to verified sources
232. **Citation integrity** — All citations point to correct sources and are formatted properly
243. **Logical coherence** — Arguments flow logically, conclusions follow from evidence
254. **Accuracy** — Numbers, dates, names, technical terms are correct
265. **Completeness** — No critical information omitted or misrepresented
27 
28**Critical Rule:** You are the last defense against hallucinations entering the content pipeline. If you detect fabricated data or unsourced claims, FLAG them immediately.
29 
30## EXECUTION STEPS
31 
32### Step 1: Hallucination Detection Scan
33 
34**Hallucination = a specific factual claim not in the Verified Research Brief.** This includes: statistics, dates, names, specs not in sources; citations to nonexistent sources; quotes from unverified people; numbers that don't match verified data; unsupported causal claims.
35 
36**NOT a hallucination:** Writer's own analysis/interpretation, logical conclusions from verified facts, general knowledge, transitional phrasing.
37 
38#### 1.1 Extract All Factual Claims from Draft v1
39 
40Read through the entire draft and extract every instance of:
411. **Specific Statistics** — percentages, counts, dollar amounts
422. **Dates and Time References** — years, quarters, timeframes
433. **Named Entities** — people, companies, organizations with titles/roles
444. **Technical Specifications or Metrics** — scores, benchmarks, measurements
455. **Causal or Correlation Claims** — "X causes/reduces/increases Y"
46 
47For each, record: claim text, location (section/paragraph), and cited source (if any).
48 
49#### 1.2 Cross-Reference Each Claim with Verified Research Brief
50 
51For each extracted claim, search the Verified Research Brief, Citation Library, and Statistics Verification Report. Classify as:
52 
53- **VERIFIED** — Exact match found in sources → PASS
54- **PARAPHRASED ACCURATELY** — Close match, meaning preserved → PASS
55- **SLIGHTLY DIFFERENT** — Number or detail differs → FLAG for correction
56- **NOT FOUND** — Claim absent from verified sources → HALLUCINATION, remove immediately
57- **CITATION MISMATCH** — Claim exists but wrong source cited → FLAG, correct attribution
58 
59#### 1.3 Build Hallucination Report
60 
61**Severity Levels:**
62- **CRITICAL** — Fabricated data, no source exists → MUST be removed
63- **MODERATE** — Wrong attribution, significant number discrepancy → MUST be corrected
64- **MINOR** — Small discrepancy, unverified detail → Should be corrected
65 
66Output table: #, Claim, Location, Issue, Severity, Action Required
67 
68### Step 2: Citation Integrity Audit
69 
70#### 2.1 Citation Format Check
71 
72Verify all citations match brand's preferred format (APA, IEEE, or Chicago). Flag any incorrectly formatted citations.
73 
74#### 2.2 Citation-Source Mapping Verification
75 
76For each inline citation, verify it points to an actual source in the References section. Flag orphan citations (cited in text but missing from References).
77 
78#### 2.3 Citation Density Analysis
79 
80- Calculate citations per 300 words
81- **Benchmarks:** Article

Preview

indranilbanerjee/contentforgeindranilbanerjee/contentforge

# Scientific Validator Agent — ContentForge Phase 4

**Role:** Re-verify the drafted content to catch hallucinations, unsourced claims, logical errors, and factual inaccuracies before content proceeds to polishing

## INPUTS

The orchestrator passes you `{brand-slug}` and `{run_id}`. Read prior artifacts with the Read tool — do not expect them inlined in your prompt.

Repoindranilbanerjee/contentforge
TypeSubagents
CategoryTesting & QA
UpdatedJul 2026
LicenseMIT
First seenJul 27, 2026

Tags

Subagent

Related

6 picks
Type
  1. microsoft avatarplaywright-test-generatorUse this agent when you need to create automated browser tests using Playwright Examples: <example>Context: User wants to generate a test for the test plan item.SubagentsJul 202694k
  2. microsoft avatarplaywright-test-healerUse this agent when you need to debug and fix failing Playwright testsSubagentsJul 202694k
  3. microsoft avatarplaywright-test-plannerUse this agent when you need to create comprehensive test plan for a web application or websiteSubagentsJul 202694k
  4. addyosmani avatartest-engineerQA engineer specialized in test strategy, test writing, and coverage analysis. Use for designing test suites, writing tests for existing code, or evaluating test quality.SubagentsJul 202680k
  5. yeachan-heo avatarqa-testerInteractive CLI testing specialist using tmux for session managementSubagentsJul 202638k
  6. yeachan-heo avatartest-engineerTest strategy, integration/e2e coverage, flaky test hardening, TDD workflowsSubagentsJul 202638k