Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add code-lawyer/Legal-Agent-Skills --skill document-fillgit clone --depth 1 https://github.com/code-lawyer/Legal-Agent-SkillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/code-lawyer/legal-agent-skills/document-fill)<a href="https://agentmods.dev/skills/code-lawyer/legal-agent-skills/document-fill"><img src="https://agentmods.dev/badge/skills/code-lawyer/legal-agent-skills/document-fill/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/code-lawyer/legal-agent-skills/document-fill"><img src="https://agentmods.dev/badge/skills/code-lawyer/legal-agent-skills/document-fill.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00061 | $0.01153 |
| Opus 5 | $0.00030 | $0.00576 |
| Sonnet 5 | $0.00012 | $0.00231 |
| Haiku 4.5 | $0.00006 | $0.00115 |
Grade A, and why
document-fill scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 10d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
文书模板填充(按本案知识库取证填空)
文件定位(先读)
本 skill 由多个文件组成。下文所有 references/…、scripts/… 路径均相对于本 SKILL.md 所在目录。开始前先确定该目录(你被指向或读取本文件时所在的位置),之后所有读取都基于它解析,不要用当前工作目录去找。
何时使用
- 用户给了一份文书模板/表单(起诉状、答辩状、合同模板等占位符文本)+ 本案知识库(案卷材料、证据、笔录等),要求按模板把待填项填好。
- 用户明确要求「按案卷填一份模板」「把这份表单填了」「用知识库填这份文书」。
- 上游技能路由到文书填充。
任务边界
本技能只做模板 × 本案知识库 的取证填充,产出干净文书 + 溯源/缺口报告。不做以下事项:
- 说理论证、法律主张的论证性文字撰写 → 路由
legal-research(论证)或起草类技能(成文) - 法律问题研究、类案检索 → 路由
legal-research - 合同审查、条款风险评估、红线稿 → 路由
contract-review
硬不变量(常驻,不可跳过)
无论任务大小、读没读全深层文件,以下底线始终生效:
- 闭世界:只使用本案材料(知识库内文件)取证;本案材料之外的任何来源(模型记忆、常识、网络)一律不得作为填入值的依据。
- 宁可查不到,也不编造:材料中确实找不到的待填项,写「未在本案材料中找到」,绝不脑补、绝不用推测值冒充事实。
- 缺口显性标注:找不到、待人工补充的待填项必须在填充计划与最终文书中显性标出(
status: gap),不得静默留白或悄悄跳过。 - 每个填入值挂且仅挂一个状态标签:
extracted/inferred/user_confirmed/ambiguous/gap/pending_drafting六选一,不得混用或省略;各状态必填字段见references/_templates/fill-plan-contract.md。 - Mode-2 强约束:
status: extracted的value必须能在其source_span.quote的原文字符串中逐字命中,不允许同义改写、数值换算或跨句拼接冒充摘录;此规则由scripts/fill_lint.py机检强制执行。 - 论证型待填项一律留空:
slot_type: argument(如"事实与理由"这类说理性文字)不属于本 skill 的填空范围,一律status: pending_drafting,交起草/论证环节处理,不由本 skill 代写。
执行顺序(按需读取,不要一次读全)
- 读
references/00-workflow.md,了解主流程与阶段划分。 - 探测/接入本案知识库时读
references/10-kb-access.md。 - 识别模板待填项、逐项取证填充时读
references/20-slot-and-fill.md。 - 出稿(渲染文书)与出溯源/缺口报告时读
references/30-output-and-report.md。
按需读取索引
references/00-workflow.md— 主流程与阶段划分references/10-kb-access.md— 本案知识库探测与接入references/20-slot-and-fill.md— 待填项识别 + 取证填充规则references/30-output-and-report.md— 出稿方式 + 溯源/缺口报告格式references/_templates/fill-plan-contract.md— 填充计划 JSON 契约(字段表 + 各 status 必填字段 + Mode-2 说明)scripts/fill_lint.py— 填充计划确定性闸门:机检来源锚定完整性与状态合法性,输出覆盖率账本scripts/fill_docx.py— 把填充计划渲染进模板;docx 优先,缺 python-docx 或模板非 docx 时降级为 markdown 并显式返回degraded=True
读取失败必须明说,不静默跳过。
What ships with it
15 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- README.md 656 B
- references/_templates/fill-plan-contract.md 4.0 KB
- references/00-workflow.md 4.1 KB
- references/10-kb-access.md 4.0 KB
- references/20-slot-and-fill.md 4.7 KB
- references/30-output-and-report.md 4.7 KB
- scripts/fill_docx.py 3.2 KB runs code
- scripts/fill_lint.py 4.7 KB runs code
- scripts/requirements.txt 111 B
- scripts/sample_fill_plan.invalid.json 824 B
- scripts/sample_fill_plan.json 1.7 KB
- scripts/test_fill_docx.py 2.3 KB runs code
- scripts/test_fill_lint.py 5.3 KB runs code
- tests/test_df_validate.py 1.8 KB runs code
- validate.py 2.9 KB runs code
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 10d ago First seen · 47 lines · 61 tokens per session scan A 35946309fb14
document-fill is a skill published in the GitHub repository code-lawyer/Legal-Agent-Skills (2 stars, last pushed 18d ago), licensed MIT. It adds 61 tokens to every session and 1,153 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
evidence-catalog-generator
Organize user-provided evidence materials into evidence catalog entries and fill them into a user-provided or generic DOCX template. The public version contains no real case template, parties, docket numbers, facts, amounts, addresses, or local file paths.
kr-patent-full-workflow
A coordinator for writing a Korean patent specification from start to finish. It guides the work through invention interviews, claim preparation, technical details, consistency checks, and final DOCX output.
kr-patent-docx-builder
Nástroj na automatické vytváranie súborov DOCX s textom patentovej prihlášky podľa formulárov Kórejského úradu pre duševné vlastníctvo.
kr-patent-format-unify
A Microsoft Word document-formatting tool for Korean patent specifications. It finds newly added paragraphs and copies the surrounding paragraph and text formatting onto them.
kr-patent-navigation-pane
A tool for adding Word document outline levels to Korean patent specifications without changing their visible formatting.
legal-citation-comprehensive
A Chinese and foreign legal-citation tool for checking, completing, explaining, and formatting footnotes under the second edition of a legal citation manual. It handles laws, cases, legal documents, and academic sources when the relevant source material is provided.