← 返回 Skills 市场
Output format (
143
总下载
0
收藏
1
当前安装
1
版本数
在 OpenClaw 中安装
/install anchor-sheet
功能描述
Extract per-subsection “anchor facts” (NO PROSE) from evidence packs so the writer is forced to include concrete numbers/benchmarks/limitations instead of ge...
使用说明 (SKILL.md)
Anchor Sheet (evidence → write hooks) [NO PROSE]
Purpose: make “what to actually say” explicit:
- select quantitative snippets (numbers/percentages)
- select evaluation anchors (benchmarks/datasets/metrics)
- select limitations/failure hooks
This prevents the writer from producing paragraph-shaped but content-poor prose.
Inputs
outline/evidence_drafts.jsonlcitations/ref.bib
Outputs
outline/anchor_sheet.jsonl
Output format (outline/anchor_sheet.jsonl)
JSONL (one object per H3 subsection).
Required fields:
sub_id,titleanchors(list; each anchor hashook_type,text,citations, and optionalpaper_id/evidence_id/pointer)
Workflow
- Read
outline/evidence_drafts.jsonl. - Prefer anchors that contain:
- a number (%, counts, scores)
- an explicit benchmark/dataset/metric name
- an explicit limitation/failure statement
- Filter anchors to only citation keys present in
citations/ref.bib. - Write
outline/anchor_sheet.jsonl.
Quality checklist
- Every H3 has >=10 cite-backed anchors (A150++ hard target).
- At least 1 anchor contains digits when the evidence pack contains digits.
- No placeholders (
TODO/…/(placeholder)).
Consumption policy (for C5 writers)
Anchors are intended to prevent “long but empty” prose. Treat them as must-use hooks, not optional ideas.
Recommended minimums per H3 (A150++):
-
=3 protocol anchors (benchmark/dataset/metric/budget/tool access)
-
=3 limitation/failure hooks (concrete, not generic “future work”)
- If digits exist in the evidence pack: include >=1 cited numeric anchor (digit + citation in the same paragraph)
Note:
- Anchor text is trimmed for readability and does not include ellipsis markers (to reduce accidental leakage into prose).
Script
Quick Start
python scripts/run.py --helppython scripts/run.py --workspace workspaces/\x3Cws>
All Options
--workspace \x3Cdir>--unit-id \x3CU###>--inputs \x3Csemicolon-separated>--outputs \x3Csemicolon-separated>--checkpoint \x3CC#>
Examples
- Default IO:
python scripts/run.py --workspace workspaces/\x3Cws>
- Explicit IO:
python scripts/run.py --workspace workspaces/\x3Cws> --inputs "outline/evidence_drafts.jsonl;citations/ref.bib" --outputs "outline/anchor_sheet.jsonl"
Refinement marker (recommended; prevents churn)
When you are satisfied with anchor facts (and they are actually subsection-specific), create:
outline/anchor_sheet.refined.ok
This is an explicit "I reviewed/refined this" signal:
- prevents scripts from regenerating and undoing your work
- (in strict runs) can be used as a completion signal before writing
安全使用建议
This skill is internally consistent: it is a Python-based extractor that reads local evidence JSONL and a BibTeX file and writes an anchor_sheet JSONL. Before installing/running, review the bundled scripts if you want to be extra cautious (they run locally and modify files under the workspace). Ensure you only point the tool at workspaces you trust (it reads and writes files there and will create a local 'refined.ok' marker to freeze results). There is no network access, no external downloads, and no secrets requested, so the primary operational risk is accidental overwriting of files in the workspace — back up any important data beforehand.
功能分析
Type: OpenClaw Skill
Name: anchor-sheet
Version: 1.0.0
The anchor-sheet skill bundle is designed to extract quantitative data, benchmarks, and limitations from evidence packs to facilitate evidence-anchored research writing. The primary execution logic in scripts/run.py and the supporting utilities in the tooling/ directory perform standard file I/O on JSONL, BibTeX, and YAML files within the workspace. No evidence of data exfiltration, malicious subprocess execution, or prompt injection was found. The code is well-structured and aligns perfectly with its stated purpose of generating structured anchor facts for downstream writing tasks.
能力评估
Purpose & Capability
Name/description say: extract numeric/benchmark/limitation anchors from evidence packs; the code and SKILL.md only require reading outline/evidence_drafts.jsonl and citations/ref.bib and writing outline/anchor_sheet.jsonl. Required binaries (python3/python) match the shipped Python scripts. No unrelated credentials or external services are requested.
Instruction Scope
Runtime instructions are narrowly scoped: read the local evidence JSONL and BibTeX, select anchors by pattern (digits, benchmark/dataset/metric keywords, limitation keywords), filter to citation keys present in the provided .bib, and write a JSONL output. The script reads/writes only workspace-local paths and enforces outline state checks; it does not access network resources or other system paths.
Install Mechanism
No install spec; this is instruction + bundled Python code. There is no remote download/install step and nothing is written outside the workspace paths the script accepts. Running requires Python available on PATH, which is proportionate.
Credentials
The skill requires no environment variables, no credentials, and no config paths. All file reads/writes are workspace-local and match the stated inputs/outputs. There are no requests for tokens/secrets or unrelated system config.
Persistence & Privilege
Flags: always:false and model invocation is allowed by default. The skill writes output and a local freeze marker (outline/anchor_sheet.refined.ok) within the workspace to prevent regeneration; it does not modify other skills or global agent config. No elevated privileges or persistent cross-skill changes are requested.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install anchor-sheet - 安装完成后，直接呼叫该 Skill 的名称或使用
/anchor-sheet触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v1.0.0
- Initial release of the anchor-sheet skill.
- Extracts concrete “anchor facts” (numbers, benchmarks, limitations) from evidence packs for each H3 subsection.
- Outputs a structured JSONL file (`outline/anchor_sheet.jsonl`) with citation-backed anchors, preventing content-poor prose.
- Emphasizes selection of facts containing numbers, benchmarks, and explicit limitations only from existing evidence.
- Ensures every H3 subsection gets at least 10 high-quality anchors, with quality and citation checks.
- Includes workflow guidance, quality checklist, minimums for anchor use, and script usage instructions.
元数据
常见问题
Anchor Sheet 是什么？
Extract per-subsection “anchor facts” (NO PROSE) from evidence packs so the writer is forced to include concrete numbers/benchmarks/limitations instead of ge... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 143 次。
如何安装 Anchor Sheet？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install anchor-sheet」即可一键安装，无需额外配置。
Anchor Sheet 是免费的吗？
是的，Anchor Sheet 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Anchor Sheet 支持哪些平台？
Anchor Sheet 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 Anchor Sheet？
由 WILLOSCAR（@willoscar）开发并维护，当前版本 v1.0.0。
推荐 Skills