Scientific Audit

Plugin sci-audit from marketplace cureonics-marketplace in mahirkurt/CureoPrivate. Commands read from the repo's manifest — 24 days ago.

Skills (7)

• ai-transparency

  Use when checking whether an LLM-generated scientific text discloses its AI use (axis F) — trigger on "is AI use disclosed", "check AI transparency", "ICMJE AI statement", "yapay zeka beyanı var mı", "AI disclosure", "undisclosed AI". Scans for a present/absent AI-use disclosure statement (ICMJE/COP

• citation-forensics

  Use when auditing the references of an LLM-generated scientific text for fabricated or hallucinated sources (axis A) — trigger on "are these citations real", "check the DOIs", "verify the references", "did the model make up this paper", "kaynak uydurma kontrolü". Resolves each DOI/PMID/arXiv id agai

• claim-grounding

  Use when checking whether the numeric and factual claims in an LLM-generated scientific text are actually supported by a cited source (axis B) — trigger on "is this number sourced", "check claim grounding", "does the citation support this", "kaynaksız iddia var mı", "iddia temellendirme". Extracts c

• hallucination-signals

  Use when scanning an LLM-generated scientific text for hallucination and confabulation signals (axis D) — trigger on "does this sound made up", "check for hallucination", "over-certainty language", "invented method names", "halüsinasyon sinyalleri", "aşırı kesinlik". Flags over-certainty absolutes,

• sci-audit-orchestrator

  Use when auditing an LLM-generated scientific text (article draft, thesis chapter, review, report, abstract) for integrity and language quality — trigger on "audit this text", "check this draft", "is this citation real", "verify these stats", "bilimsel metni denetle", "kaynakları doğrula", "Türkçe i

• stats-forensics

  Use when checking the internal statistical consistency of an LLM-generated scientific text (axis C) — trigger on "check the stats", "do these p-values add up", "statcheck this", "GRIM test", "are these percentages consistent", "istatistik tutarlılığı", "p değeri kontrolü". Recomputes p-values from r

• turkish-sci-style

  Use when auditing Turkish scientific text for writing quality and orthography (axis G) — auto-triggers when the text is Turkish or the user passes --lang tr. Trigger on "Türkçe imla denetimi", "bilimsel Türkçe kontrol", "ondalık virgül", "akademik dil denetimi", "check Turkish scientific writing", "

Slash commands (7)

• /sci-audit:ai-log

  Append an audit-run entry to the sci-audit JSONL log

• /sci-audit:audit-report

  Compile per-axis findings into a single merged sci-audit report

• /sci-audit:audit

  Full seven-axis scientific-integrity audit of an LLM-generated text

• /sci-audit:check-stats

  Axis C only — recompute p-values, run GRIM, flag impossible statistics

• /sci-audit:check-turkish

  Axis G solo — Turkish scientific writing & orthography audit

• /sci-audit:guideline-check

  Axis E only — check a manuscript against a reporting guideline

• /sci-audit:verify-citations

  Axis A only — detect fabricated, misattributed, or retracted references

Subagents (8)

• citation-verifier

  Resolves individual scientific references (DOI/PMID/arXiv id/title) against scholarly MCP servers and reports whether each exists, whether its metadata matches what the text claims, and whether it is retracted. Use for axis A (reference integrity).

• claim-extractor

  Converts a section of scientific text into a structured list of discrete, checkable claims (each with its quantitative content and any cited source), so downstream verification can test them one by one. Use for axis B (claim grounding).

• claim-refuter

  Adversarially tests whether a claim's cited source actually supports it — defaulting to "unsupported" unless the source clearly backs the claim. Use for axis B (claim grounding) after claims are extracted. Resolves sources via scholarly MCPs.

• entity-verifier

  Verifies that named biomedical/scientific entities in a text (drugs, active ingredients, genes, diseases, institutions) actually exist by resolving them against authority terminology MCPs. Use for axis D to catch hallucinated entity names (a fabricated drug or gene is catchable where a fabricated se

• entropy-sampler

  Drives the semantic-entropy confabulation signal for a single high-stakes factual claim — samples several independent answers to the same underlying question and clusters them by meaning. Use for axis D when a specific claim needs a consistency check beyond the deterministic signal scan.

• guideline-mapper

  Maps a manuscript against a reporting-guideline checklist (PRISMA/CONSORT/STROBE/COREQ/SRQR/JARS/TRIPOD), marking each item present/missing/partial with the locating quote. Use for axis E (reporting-guideline conformance).

• stats-checker

  Runs the deterministic statistics-forensics script and adds domain reasoning about effect-size and heterogeneity plausibility on top of its arithmetic output. Use for axis C (statistical consistency).

• style-judge

  Claude-native Turkish scientific-writing judge for register, fluency, causal-language discipline, and terminology consistency (axis G LLM layer). Applies the rubric that the deterministic tr_sciaudit core cannot; deterministic findings win on conflict. Use only when the text is Turkish.

MCP servers (4)

• pubmed

  connects to: https://pubmed.mcp.claude.com/mcp

• pubmed-epmc

  connects to: https://pubmed.cureonics.com/mcp

• openalex

  connects to: https://openalex.cureonics.com/mcp

• semantic-scholar

  connects to: https://semanticscholar.cureonics.com/mcp

Hooks (5)

• SessionStart

  /usr/bin/env python3 "${CLAUDE_PLUGIN_ROOT}/hooks/scripts/session_start.py"

• UserPromptSubmit

  /usr/bin/env python3 "${CLAUDE_PLUGIN_ROOT}/hooks/scripts/user_prompt_submit.py"

• PreToolUse

  /usr/bin/env python3 "${CLAUDE_PLUGIN_ROOT}/hooks/scripts/pre_tool_use_policy.py"

• PostToolUse

  /usr/bin/env python3 "${CLAUDE_PLUGIN_ROOT}/hooks/scripts/post_tool_use_review.py"

• Stop

  /usr/bin/env python3 "${CLAUDE_PLUGIN_ROOT}/hooks/scripts/stop_verify.py"

Similar plugins