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 rojim666/SztuCode --skill general-writergit clone --depth 1 https://github.com/rojim666/SztuCodeWrote 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/rojim666/sztucode/general-writer)<a href="https://agentmods.dev/skills/rojim666/sztucode/general-writer"><img src="https://agentmods.dev/badge/skills/rojim666/sztucode/general-writer/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/rojim666/sztucode/general-writer"><img src="https://agentmods.dev/badge/skills/rojim666/sztucode/general-writer.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 | $0.00115 | $0.01645 |
| Opus 5.5 | $0.00046 | $0.00658 |
| Sonnet 5.5 | $0.00023 | $0.00329 |
| Haiku 4.5 | $0.00012 | $0.00164 |
Grade A, and why
general-writer 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 1mo 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.
How it starts
The opening of the file, as written. The whole thing — 157 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Phase 1 — 主题理解
你需要详细思考如下要点:
- 文体归类(正式公文?新媒体?内部周报?)→ 查询
experts/general-writer/references/doc-type-matrix.md匹配字数与风格。 - 目标读者画像(同事?客户?公众?)。
- 核心信息点 3–5 条(用户已提供的 + 需要检索补全的)。
保存到
output/params/topic.yamldetected_doc_type: "工作周报" target_length: 1000 target_tone: "第一人称,正式度★★★☆☆"
Phase 2 — Research(广度优先)
如果判断完全不需要研究时(纯格式化、纯模板填空),直接跳过,不需要传空参数;否则任何先充分了解再动笔"的阶段都不能跳过本阶段,典型场景包括:
- 写作前的素材准备
- 大纲生成前的行业背景摸底
- 针对某章节的信息缺口补充
你需要先参考参数字典
<plugin_root>/core/engines/deep-research/README.md生成驱动引擎的参数,保存到output/params/deep-research.yaml。生成参数文件后,再读取<plugin_root>/core/engines/deep-research/engine.md执行 6 步状态机,产出信息库快照到snapshot_path。
默认deepredearch参数如下仅供参考:
# 必填参数(运行时由 Phase 1 结果填入)
topic: "<由 Phase 1 主题理解结果填入,一句话描述>"
tools: ["web_search"] # L1 默认仅用网络检索
# 可选参数(已为 L1 通用写作配置默认值)
scope: "web" # 广度优先,可上网检索
max_loops: 2 # 通用写作不需要过深研究
min_facts_required: 5 # 累计结构化事实数 ≥ 5
gap_check_strict: "relaxed" # 宽松模式:Q1-Q4 必须通过,Q5-Q6 可容忍
search_dimensions: # 必搜维度,默认通用两维
- "数据"
- "趋势"
tier_priority: [] # L1 不限定权威源优先级
snapshot_path: "output/research/{topic}_research.md"
Phase 3 — Writing(纯写作 + 输出 critic_config)
Step 3.0 — 生成 critic_config
根据 Phase 1 产出的 output/params/topic.yaml,生成 critic_config 声明(由 doc-writer 消费):
# ===== 模式提示 =====
critic_mode_hint: "once" # L1 默认 once;target_length ≥ 3000 时建议 per-section
force_mode: null # L1 不强制,允许 skip
# ===== 评分标准 =====
rubrics_files:
- "experts/general-writer/references/quality-framework.md"
- "experts/general-writer/references/doc-type-matrix.md"
pass_score: 75
max_loops: 2
# ===== per-section 模式参数 =====
section_params:
scope: "section"
min_issues: 1
min_instructions: 2
# ===== full 模式参数 =====
full_params:
scope: "full"
min_issues: 2
min_instructions: 3
# ===== revision 模式参数 =====
revision_params:
scope: "revision"
min_issues: 1
min_instructions: 1
What ships with it
6 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.
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.
- 1mo ago First seen · 157 lines · 115 tokens per session scan A 567043caf2cc
general-writer is a skill published in the GitHub repository rojim666/SztuCode (110 stars, last pushed yesterday), licensed MIT. It adds 115 tokens to every session and 1,645 once invoked, about $0.0005 per session on Opus 5.5. 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 2026-09-12.
Other skills, from other repositories
pptx-official
Creates, edits, inspects, and renders PowerPoint presentations (.pptx, .potx): pitch decks, executive readouts, board and training material, template fills, slide reordering or duplication, speaker notes, text and chart-data extraction, and deck-to-PDF or slide-to-image rendering. Use when a PowerPoint file is the…
docx-official
Creates, edits, inspects, and converts Microsoft Word documents (.docx, .dotx): reports, letters, memos, contracts, proposals, filled templates, tracked-change acceptance, review comments, and text, table, or metadata extraction. Use when a Word file is the deliverable or the source of record, or the user mentions…
xlsx-official
Builds, edits, cleans, and inspects spreadsheet files (.xlsx, .xltx, .csv, .tsv) when the file is the deliverable or the record being changed: financial and operating models with live formulas, formula-driven summaries, template fills, cell and row patches, messy-data repair, sheet-to-CSV or PDF export, formula…
pdf-official
Extracts text, tables, metadata, and images from PDF files; merges, splits, rotates, crops, watermarks, encrypts, compresses, and repairs them; composes new PDFs such as reports, invoices, and certificates; fills AcroForm fields or overlays values on non-fillable and scanned forms; runs OCR on scanned pages; renders…
office-annotation-edit
Applies targeted edits to the parts of a PowerPoint deck or Excel workbook a user selected in the EvoFlux document viewer: a message carrying an evoflux-annotations block names the file, the slide and its selected shapes or area, or the sheet and its selected cell range, and an instruction per annotation. Use whenever…
pptx
Create/edit/inspect/verify slide decks and PPTX presentations.