跳到正文
cccskills
免费在 GitHub 公开

module-source-coverage

Use when a tutorial module plan exists and the run needs an auditable module-to-source grounding file before prose.

中文简介正在准备中。当前显示原文说明。

查看安装方法

包含的文件(2)

  • SKILL.md1.9 KB
  • scripts/run.py1.4 KB

SKILL.md(原文)

安装前可以查看交给 AI 代理的指令内容。

Module Source Coverage

Triggers & routing

  • Trigger: module coverage, source coverage, tutorial grounding, 模块覆盖, 来源覆盖.
  • Use when: source-tutorial 的 C2,已有 outline/module_plan.yml,需要确认每个模块都能回指到 sources。
  • Skip if: module plan 或 source ingest 不完整。
  • Network: none.
  • Guardrail: 只做 grounding audit,不写教程正文。

Builds outline/source_coverage.jsonl, one coverage record per module.

Inputs

  • outline/module_plan.yml
  • sources/index.jsonl
  • sources/provenance.jsonl

Output

  • outline/source_coverage.jsonl

Contract

Each module record must include:

  • module_id
  • module_title
  • source_ids and/or explicit gaps

Plus exactly one corpus-reconciliation record (record_type: corpus_reconciliation, no module_id) listing ingested_source_ids, attributed_source_ids, and unused_source_ids, with an explicit gaps entry whenever a source was ingested but used by no module. A per-module audit alone cannot surface a source that contributed to zero modules, so this record keeps an unused source explicit instead of silently dropped.

Script boundary

scripts/run.py should:

  • score module-to-source relevance
  • choose a small source set per module
  • record explicit grounding gaps when coverage is weak
  • append the corpus-reconciliation record so unused ingested sources are flagged

Keep text matching and snippet scoring in shared tutorial tooling, not in the wrapper.

Acceptance

  • outline/source_coverage.jsonl exists
  • every module appears exactly once
  • exactly one corpus_reconciliation record accounts for every ingested source
  • missing grounding is explicit instead of silent

Non-goals

  • context pack assembly
  • tutorial prose

评价

还没有评价。欢迎分享你的使用感受。

评论

无需登录即可发表。欢迎分享使用感受、提问,或说明在哪些环境下无法运行。

还没有评论。来写第一条吧。

提交后会立即公开。请勿填写邮箱、电话等个人信息。URL 最多 2 个。

0 / 1000 字

此表单受 reCAPTCHA 保护,适用 Google 的隐私权政策和服务条款。

同一仓库的技能

Internal template for creating or refactoring a skill into the repository's reference-first shape. **Trigger**: reference-first template, blueprint skill, create a reusable skill, refactor a script-heavy skill. **Use when**: you need a lean `SKILL.md`, explicit `references/`, machine-readable `assets/`, and a minimal deterministic `run.py`.

中文简介正在准备中。当前显示原文说明。

WILLOSCAR/research-units-pipeline-skills5132026年10月5日 更新

Download a small corpus of open-access arXiv survey/review PDFs about agentic systems and extract text for style learning.

中文简介正在准备中。当前显示原文说明。

WILLOSCAR/research-units-pipeline-skills5132026年10月5日 更新

Extract per-subsection “anchor facts” (NO PROSE) from evidence packs so the writer is forced to include concrete numbers/benchmarks/limitations instead of generic summaries.

中文简介正在准备中。当前显示原文说明。

WILLOSCAR/research-units-pipeline-skills5132026年10月5日 更新

Curate reader-facing survey tables for the Appendix (clean layout + high information density), using only in-scope evidence and existing citation keys.

中文简介正在准备中。当前显示原文说明。

WILLOSCAR/research-units-pipeline-skills5132026年10月5日 更新

Build a final, machine-readable argument snapshot from current H3 sections and refresh section fingerprints before merge.

中文简介正在准备中。当前显示原文说明。

WILLOSCAR/research-units-pipeline-skills5132026年10月5日 更新

Audit one research Workspace for declared Unit outputs and Pipeline target Artifacts, writing `output/CONTRACT_REPORT.md`; use for mid-Run coverage snapshots or final delivery completeness, not deep provenance integrity.

中文简介正在准备中。当前显示原文说明。

WILLOSCAR/research-units-pipeline-skills5132026年10月5日 更新

查看 WILLOSCAR 的全部技能

报告此技能的问题