validate-research-gap

使い方

インストールコマンドをコピーし、自分の環境へ導入する前にソースを確認します。
必要なツール、認証情報、対応環境を原文で確認します。
入力と成果物を指定し、出力を資料と照合します。

npx skills add https://github.com/yogsoth-ai/de-anthropocentric-research-engine --skill validate-research-gap

開始前の確認

導入にはモデル契約、API枠、ファイルへのアクセス許可は含まれません。
スター数や導入数は活動の指標であり、独立した品質・安全性試験ではありません。

リポジトリの原文

見出し、例、コード、表、リンク、参照画像を含む原文を表示しています。

Purpose

Test whether an apparent research gap is real, persistent, and not an evidence-acquisition artifact, and keep the candidates rejected along the way auditable.

Input contract

yaml

required: [gap_claim, evidence_set, search_scope]
optional: [independent_sources, time_slices, candidate_gap_types]
constraints: [absence claims require an explicit scope and retrieval record]

Execution protocol

Do not perform called SOP operations inline; each loaded SOP owns its contract and thresholds.

You MUST load skill verify-evidence-independence to check source independence.
You MUST load skill filter-false-gap to separate open, touched-but-not-closed, and false gaps.
You MUST load skill analyze-temporal-trajectory to test persistence across time.
You MUST load skill classify-research-gap to classify each validated gap, including those prior work touches without closing.

If several validated gaps require comparative priority, consider rank-candidates. If stakeholder boundaries determine whether the gap matters, consider map-stakeholder-system. Deviation: if temporal data are unavailable, return persistence unknown rather than infer stability.

Output contract

yaml

produces: [gap_validity_verdict, independence_check, persistence_assessment, typed_gap, rejection_register]
delta_fields: [findings, evidence_updates, uncertainties, decisions, recommended_jumps]

Thresholds and quality gates

B: independence, search adequacy, persistence, and gap type are separately reported; unsearched space is not a confirmed gap.
B: every candidate rejected as resolved appears in the rejection register with its closing work; a rejection resting on a single source is flagged.
When the typology admits contradiction, measurement, or boundary gaps and none is reported, the verdict states why.

Failure and counterexamples

Reject gaps caused by duplicate sources, narrow retrieval, prior work that closes the question, or unanswerable wording. Do not reject a gap because prior work exists in its region.