Agent skill

Blog 3p Harness

by flaqai in flaqai/backlink_skills

初始化、冻结、恢复并机械校验仅人工原生发布交接的博客 3P 内容项目;要求用户先确认完整写前方案和每篇平台/账号映射,严格限制用户提供的平台范围,且不自动发布。用于启动内容项目、准备工作区、汇报写前方案、锁定需求或校验工作流结构。

MITAuto-check passed

Install Blog 3p Harness

skills CLI
$ npx skills add flaqai/backlink_skills --skill blog-3p-harness -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install flaqai/backlink_skills blog-3p-harness --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/flaqai/backlink_skills.git skills-src && mkdir -p .claude/skills && cp -r skills-src/blog-3p-skill-matrix/skills/blog-3p-harness .claude/skills/blog-3p-harness && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
blog-3p-harness
GitHub stars
756
Token cost
~1.5k tokens
SKILL.md length
285 words
Files
6 (incl. scripts, references)
Skills in repo
19
Repo updated
First seen
Licence
MIT

At a glance

初始化、冻结、恢复并机械校验仅人工原生发布交接的博客 3P 内容项目;要求用户先确认完整写前方案和每篇平台/账号映射,严格限制用户提供的平台范围,且不自动发布。用于启动内容项目、准备工作区、汇报写前方案、锁定需求或校验工作流结构。

  • Works in 4 steps: G 完成写前方案、保存并绑定用户确认回执后,运行 check 和… → blog-3p-writer 是唯一可执行… → blog-3p-gate 在一个 BATCH_GATE_ACCEPTANCE… → …
  • SKILL.md covers 初始化, 写前方案与用户确认, 平台范围与本地化 and 运行与交接, plus 1 more section
  • Runs Python scripts from its folder

What it does

Blog 3p Harness is an agent skill from flaqai/backlink_skills. 初始化、冻结、恢复并机械校验仅人工原生发布交接的博客 3P 内容项目;要求用户先确认完整写前方案和每篇平台/账号映射,严格限制用户提供的平台范围,且不自动发布。用于启动内容项目、准备工作区、汇报写前方案、锁定需求或校验工作流结构。

Its SKILL.md is about 1.5k tokens, which your agent loads only when the skill is triggered. The skill folder holds 8 other files, including scripts and reference files (for example `agents/openai.yaml`, `references/campaign-contract.md` and `references/pre-clearance.md`).

The repository describes itself as: Awesome skills for submitting url to free websites. Get more backlinks for your website to get more traffic. The licence is MIT.

Example prompts

  • “/blog-3p-harness”

Requirements

  • Python 3

Workflow steps

4 steps, taken from the first numbered list in SKILL.md.

  1. G 完成写前方案、保存并绑定用户确认回执后,运行 check 和 dispatch-readiness。前者仅证明结构可读;后者必须显示 DISPATCH_READY,否则先补齐明确列出的 REQ-*、可见项目 G 登记、每篇人工发布映射或语言/平台兼容记录,再创建…
  2. blog-3p-writer 是唯一可执行 W;blog-writer-merged 只作编辑参考。W 连续完成研究、成稿、图片和双载荷,再以 CANDIDATE_LOCK → ADVERSARIAL_SELF_QA → TARGETED_REPAIR →…
  3. blog-3p-gate 在一个 BATCH_GATE_ACCEPTANCE 回合中验收当前所有具备适用质量凭据的文章:WQ 是 W 的可见 AUTHOR_QA_READY 哈希回执,WR 是独立 R 的 APPROVED;升级 WQ 必须有后者。每篇保留独立结论行。W…
  4. 人工回传 HUMAN_ACCEPTED URL 后,校验回执并生成只读快照;复用同一项目统筹 G 的 PUBLIC_QA_BATCH_READONLY 回合比对当前已回传文章,不重启 W/R。

What it can do on your machine

Read from SKILL.md and the folder at commit 3c56c94. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Ships 1 file in scripts/ (Python), which the agent can run.

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Blog 3p Harness loads about 1.5k tokens when it runs, and up to ~11k if it reads all its reference files. Until then it costs about 33 tokens; SKILL.md has 285 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~33
When it runs · the whole SKILL.md, loaded when a task matches
~1.5k
With references · SKILL.md plus every file in references/, read only if the agent opens them
~11k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); the scripts in this folder are not scanned.

SKILL.md

The full file from flaqai/backlink_skills at commit 3c56c94, republished under its MIT licence (© flaqai). 285 words, ~1,462 tokens.

Download SKILL.mdSave it as .claude/skills/blog-3p-harness/SKILL.md (or your agent's skills folder). This skill also uses 5 other files; get the full folder from GitHub.
name
blog-3p-harness
description
初始化、冻结、恢复并机械校验仅人工原生发布交接的博客 3P 内容项目;要求用户先确认完整写前方案和每篇平台/账号映射,严格限制用户提供的平台范围,且不自动发布。用于启动内容项目、准备工作区、汇报写前方案、锁定需求或校验工作流结构。

博客 3P:本地流程工具

先阅读术语约定。本技能负责可恢复的本地工作区和机械校验,不代替编辑判断、平台发布或人工原生编辑器操作。

初始化

运行 scripts/harnessctl.py init --workspace <absolute-path> --campaign-id <id>。它只创建本地文件:campaign.json、state.json、confirmation.md、requirements-contract.md、唯一可编辑的 prewrite-plan.json、确定性只读视图 prewrite-plan.md、一次性 pre-clearance-checklist.md 及标准证据/输出目录。当前 schema-2.14 工作区使用写前方案 schema 1.8、evidence pack schema 1.3 和 article-package.json@1.5,默认持续主会话及 articles/<article_id>/{context,research,canonical,reviews,handoff} 路径隔离,并保存原始用户确认回执;schema-2.13 的独立 R 及更早版本按各自历史规则读取/校验,不自动转成 WQ。当前工作区只支持人工原生发布交接;脚本绝不执行平台写入。

写前方案与用户确认

G 的第一阶段是 prewrite_planning。在用户确认前,不得存在文章隔离路径、W/R 派发、文章队列、大纲、正文、图片、唯一基准稿或文章调研子角色。G 编辑 prewrite-plan.json 后运行 sync-prewrite-plan,再向用户展示生成的 prewrite-plan.md;Markdown 是只读视图,禁止双份手填。

每篇方案卡固定为四个可读区块:任务/受众;独立读者价值和必保留次要 CTA;editorial_brief;风险与待用户决定事项。还须展示 frozen_delivery_mapping、topic_slot(读者任务、核心意图、市场、差异化角度、禁止偏离项)和 evidence_posture。topic_slot 不锁字面关键词/标题。editorial_brief 汇总关键词/本地化路径、事实边界、暂定结构、图文叙事与人工传递假设,不双份维护研究。只展示实际共享证据。冻结 review_effort:默认 AUTHOR_QA_INTEGRATED / WQ_SHARED_CONTEXT / INTEGRATED_IN_AUTHOR_QA;用户要求或具体风险时用 INDEPENDENT_R_ESCALATION / WR_INDEPENDENT,仅 WR 内确有成文前风险才追加独立研究审。不可证实的内容标 UNVERIFIED 并说明风险;写前方案不等于研究批准,W 仍要一体化研究。

只有明确的 OWNER_PREWRITE_PLAN_CONFIRMED 才允许启动文章任务。当前 schema 2.14 须先保存用户确认原文到 evidence/owner-confirmations/,再运行 confirm-prewrite-plan --confirmation-id ... --receipt-file ... --receipt-type OWNER_MESSAGE|OWNER_FILE --source-locator ...。该命令绑定文章集合、计划/回执哈希、受保护范围及来源/时间;手工改状态不能提升方案。语言、市场、平台、账号、适配模式、跨语言例外、topic_slot 或写作姿态变化时,sync-prewrite-plan 返回 SCOPE_RECONFIRM_REQUIRED 且保留旧回执;须显式失效后取得新确认。OWNER_PREWRITE_PLAN_CHANGES_REQUESTED 不允许部分启动。

平台范围与本地化

  • 没有平台映射时,G 只可创建或恢复一个可见、可复用的 blog-3p-platform-matching 子角色;它只在用户提供的 XLSX/表格/消息候选中做语言/市场/内容格式匹配。G 只核对来源与唯一性,不排序、不选平台。
  • 当前 schema 2.14 只使用 HUMAN_RELEASE_ONLY_V1 与 release_policy.mode = HUMAN_NATIVE_ONLY;machine_external_writes_allowed = false。human_release_requested 不能作为跳过人工发布映射的开关。platform_scope.allowed_pairs 仅投影已确认的用户平台选择,每篇在 WQ/WR 派发前须有不同 {platform, account} 与兼容主读者记录;pair 的用户来源哈希、准确定位、平台原文、账号确认原文和确认 ID 缺一不可。
  • 登录页、注册公告、编辑器行为、会话恢复、内部配置和历史记录只能说明可承接性,绝不能新增允许的平台/账号。模糊的“新增语言”或“平台不重复”请求没有平台选择授权,只能返回 OWNER_DECISION_REQUIRED。
  • 文章语言来自用户确认。每篇 locale_platform_validation 必须分开记录 primary_reader_languages、primary_reader_markets、primary_audience_evidence_path 与 transport_supported_content_languages、transport_evidence_path。平台支持语言只决定传递资格,不能证明英语或任何语言是平台主读者语言;默认只有 PRIMARY_AUDIENCE_MATCH 可派发。语言/市场不匹配只能是逐篇、逐字绑定的 CROSS_LANGUAGE_EXCEPTION_OWNER_CONFIRMED,否则写 PLATFORM_AUDIENCE_MISMATCH_RECONFIRM_OWNER,不得自动翻译或改配。
  • 跨语言用语顺序固定为 CURRENT_BRAND_SITE → REGIONAL_SERP → MODEL_TRANSLATION_FALLBACK。Google Trends 如可用,只比较英文种子和英文候选,并且只是全球相对关注度背景;数据不足、无 Breakout 或无清晰趋势时,不得强求或阻塞,继续以当前品牌站与地区 SERP 的有边界证据选择长尾意图。两次地区 SERP 核验都没有可用共识时,才以有据模型翻译兜底。平台资料只能决定表达形式、技术深度和传递风险,不能作为关键词、自然变体、搜索意图、当地需求或热门度依据。

运行与交接

  1. G 完成写前方案、保存并绑定用户确认回执后,运行 check 和 dispatch-readiness。前者仅证明结构可读;后者必须显示 DISPATCH_READY,否则先补齐明确列出的 REQ-*、可见项目 G 登记、每篇人工发布映射或语言/平台兼容记录,再创建 articles/<article_id>/ 的隔离交付路径并派发可见 W/R。build-article-context 的 --workspace 始终是含 campaign.json、state.json、prewrite-plan.json 与 requirements-contract.md 的内容项目根目录;文章产物由其下的确定性 articles/<article_id>/ 路径承接,不能使用新建的空目录作为内容项目根。
  2. blog-3p-writer 是唯一可执行 W;blog-writer-merged 只作编辑参考。W 连续完成研究、成稿、图片和双载荷,再以 CANDIDATE_LOCK → ADVERSARIAL_SELF_QA → TARGETED_REPAIR → FINAL_BIND 自查本篇证据、正文、CTA、图文语义与载荷并绑定最终哈希。默认 WQ 只可登记 AUTHOR_QA_READY,不生成独立 R 的 APPROVED。用户要求或记录高风险时走 WR,独立 R 一次 FULL_REVIEW,仅成文前研究风险另加研究审。机械投影校验正常只语义读 HTML,COMPANION_DUAL_READ_REQUIRED 才双读 Markdown。
  3. blog-3p-gate 在一个 BATCH_GATE_ACCEPTANCE 回合中验收当前所有具备适用质量凭据的文章:WQ 是 W 的可见 AUTHOR_QA_READY 哈希回执,WR 是独立 R 的 APPROVED;升级 WQ 必须有后者。每篇保留独立结论行。W 在质量核验前编译交付页;blog-3p-human-handoff 仅在 HUMAN_RELEASE_READY 后生成派生清单,不改写已核验载荷。
  4. 人工回传 HUMAN_ACCEPTED URL 后,校验回执并生成只读快照;复用同一项目统筹 G 的 PUBLIC_QA_BATCH_READONLY 回合比对当前已回传文章,不重启 W/R。

可见角色会话、文章路径隔离与 Git 工作树是不同概念。默认持续主会话和文章路径隔离:可见 W 可跨普通 WQ 文章复用;WR 才激活独立于 W 的 R。每篇只读写自己的 articles/<article_id>/;全项目只登记一个 G。Git 工作树只在记录真实并发写入、高风险重写/回滚或用户要求 Git 隔离时允许,槽位或 token 节省不是理由。CLI 只可执行确定性本地操作,不能冒充可见 WQ/WR 质量任务。

每次交接前运行 scripts/harnessctl.py check --workspace <campaign-root>。在 WQ 最终自查或 WR FULL_REVIEW 前,W 以内容项目根运行 check-review-ready --workspace <campaign-root> --article-contract articles/<article_id>/context/article-contract.json --review-index articles/<article_id>/reviews/review-index.json --article-package articles/<article_id>/article-package.json。它只校验身份、哈希、来源引用结构、固定载荷、冻结 CTA 与伴随投影,不作质量批准。TOPIC_EVIDENCE_CONFLICT 须同一 G 的有据决策后才能恢复。当前 schema-2.14 再运行 check-article-package、check-handoff-manifest、G 批量 check-batch-gate;它们验证适用质量凭据:WQ AUTHOR_QA_READY、WR 独立 APPROVED,不评分文章。WQ 改稿/改图/改载荷重跑同一 W 的完整自查并换新回执,不走 R-Δ;WR 的有限修复才可由同一 R 作有边界 R-Δ。公开回传先校验 receipt。

需要量化而不是猜测时,运行 summarize-efficiency --workspace <campaign-workspace> [--output evidence/operations/efficiency-summary.json]。它只读取已有任务记录;input_tokens、output_tokens、wall_time_seconds 和 tool_calls 都是可选运行时字段,缺失时明确输出 UNAVAILABLE,不写入文章包、不成为发布门,也不启动模型或新角色。

恢复

发生压缩、恢复或上下文不确定时,先重读流程核心,再读取 campaign.json、state.json、requirements-contract.md、当前文章的 articles/<article_id>/context/article-contract.json 与 articles/<article_id>/reviews/review-index.json,以及当前变更文件。只有哈希漂移、问题项谱系不清、代理替换或升级时,才完整重读历史报告;在状态历史中追加 CONTEXT_REHYDRATED,不得推断缺失证据。

日常先以流程核心和当前结构化记录行动;仅在范围冻结、恢复异常或对应风险触发时,按需读取内容项目契约或写前检查清单的相关段落。机器字段、状态码、命令参数与路径保持不变。

© flaqai, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 5 other files (scripts, references) in blog-3p-skill-matrix/skills/blog-3p-harness of flaqai/backlink_skills.

  • SKILL.md
  • agents/openai.yaml
  • references/campaign-contract.md
  • references/pre-clearance.md
  • references/shared-evidence-cache.md
  • scripts/harnessctl.py

Open the folder on GitHubat commit 3c56c94

Compare with similar skills

Blog 3p Harness next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Blog 3p Harness compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Blog 3p Harness this skillflaqai/backlink_skills756—~1.5kAutomated safety check: PassMIT
BlogAgriciDaniel/claude-blog2.3k—~6.2kAutomated safety check: PassMIT
BlogAgriciDaniel/claude-blog2.3k1 repos~8.6kAutomated safety check: WarnMIT
Notion To Blogwasp-lang/wasp19k—~922Automated safety check: PassMIT
Blog Writing Guidesickn33/agentic-awesome-skills47k2 repos~2.2kAutomated safety check: PassMIT
Blog Post Drafterluongnv89/claude-howto42k—~2.2kAutomated safety check: PassMIT

Similar skills

  • Blog

    AgriciDaniel/claude-blog

    Full-lifecycle blog engine with 31 sub-skills, 12 templates, 100-point scoring, and 5 agents.

    2.3k GitHub stars~6.2k tokensUpdated 2 days ago
    Writing & ContentAuto-check passed
  • Blog

    AgriciDaniel/claude-blog

    Full-lifecycle blog engine with 31 sub-skills, 12 content templates, 5-category 100-point scoring, and 5 specialized agents.

    2.3k GitHub starsUsed in 1 repo~8.6k tokens
    Writing & ContentAuto-check: warnings
  • Notion To Blog

    wasp-lang/wasp

    Transfer a blog post from Notion to the Wasp blog. An agent skill from wasp-lang/wasp.

    19k GitHub stars~922 tokensUpdated today
    Writing & ContentAuto-check passed
  • Blog Writing Guide

    sickn33/agentic-awesome-skills

    This skill enforces Sentry's blog writing standards across every post — whether you're helping an engineer write their first blog post or a marketer draft a product announcement.

    47k GitHub starsUsed in 2 repos~2.2k tokens
    Writing & ContentAuto-check passed
  • Blog Post Drafter

    luongnv89/claude-howto

    Guides a blog post from idea to draft in stages: research notes from your sources, brainstorming and clarifying questions, then outlining and versioned drafting.

    42k GitHub stars~2.2k tokensUpdated today
    Writing & ContentAuto-check passed
  • Blog Draft Workflow

    luongnv89/claude-howto

    Walks through writing a blog post from an idea and source material: research, brainstorming, outline approval, versioned drafts and optional git commits.

    42k GitHub stars~1.1k tokensUpdated today
    Writing & ContentAuto-check passed

More from flaqai/backlink_skills

All 19 skills in this repo
  • SPD V1 Batch. An agent skill from flaqai/backlink_skills.

    756 GitHub stars~1.7k tokensUpdated 17 days ago
    Auto-check passed
  • Wechat Writer

    flaqai/backlink_skills

    微信公众号文章的选题、研究、写作、改稿、审稿与交付工作流。Use when creating, outlining, researching, rewriting, auditing, or packaging Chinese WeChat Official Account articles, 微信文章, 公众号推文, 公众号长文, 热点解读, 干货教程, 观点文, 故事文, 对比测评…

    756 GitHub stars~1.8k tokensUpdated 17 days ago
    Auto-check passed
  • Backlink Batch Expansion

    flaqai/backlink_skills

    Build and maintain a reusable backlink-submission candidate pool with global deduplication, fee and brand classification, configurable batch exports, and executor feedback.

    756 GitHub stars~757 tokensUpdated 17 days ago
    Auto-check passed
  • Linkedin Writer

    flaqai/backlink_skills

    LinkedIn-native long-form article and newsletter writing workflow for LinkedIn and Google-to-LinkedIn topic discovery, business-depth research, professional thought leadership, evidence-led…

    756 GitHub stars~6k tokensUpdated 17 days ago
    Auto-check passed
  • Medium Writer

    flaqai/backlink_skills

    Medium-native long-form research, author-assistance, writing, editing, review, topic discovery, publication matching, and packaging workflow.

    756 GitHub stars~5.6k tokensUpdated 17 days ago
    Auto-check passed

Questions about Blog 3p Harness

What does Blog 3p Harness do?

初始化、冻结、恢复并机械校验仅人工原生发布交接的博客 3P 内容项目;要求用户先确认完整写前方案和每篇平台/账号映射,严格限制用户提供的平台范围,且不自动发布。用于启动内容项目、准备工作区、汇报写前方案、锁定需求或校验工作流结构。. Blog 3p Harness is an agent skill from flaqai/backlink_skills.

How do I install Blog 3p Harness in Claude Code?

Run `npx skills add flaqai/backlink_skills --skill blog-3p-harness -a claude-code`. Or copy the skill folder (blog-3p-skill-matrix/skills/blog-3p-harness in flaqai/backlink_skills) into .claude/skills/blog-3p-harness in your project. Claude Code loads it when a task matches its description.

How do I install Blog 3p Harness in Codex?

Run `npx skills add flaqai/backlink_skills --skill blog-3p-harness -a codex`. Or copy the skill folder (blog-3p-skill-matrix/skills/blog-3p-harness in flaqai/backlink_skills) into .agents/skills/blog-3p-harness in your project. Codex loads it when a task matches its description.

Can I use Blog 3p Harness in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add flaqai/backlink_skills --skill blog-3p-harness -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/blog-3p-harness, .gemini/skills/blog-3p-harness, .github/skills/blog-3p-harness and .opencode/skills/blog-3p-harness in your project.

What does Blog 3p Harness need to run?

Going by SKILL.md and its folder, Blog 3p Harness needs Python for the scripts in its folder. Our summary lists: Python 3.

Does Blog 3p Harness access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Blog 3p Harness safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. The check reads SKILL.md only: the scripts in the folder are not scanned, so read them before running anything.

What licence does Blog 3p Harness use?

Blog 3p Harness is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Blog 3p Harness use?

About 1.5k tokens (SKILL.md is roughly 5.8k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 9.2k tokens, read only when the agent opens those files.

What are the alternatives to Blog 3p Harness?

Skills that share tags, products or a category with Blog 3p Harness: Blog (AgriciDaniel/claude-blog, 2.3k stars), Blog (AgriciDaniel/claude-blog, 2.3k stars), Notion To Blog (wasp-lang/wasp, 19k stars) and Blog Writing Guide (sickn33/agentic-awesome-skills, 47k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Blog 3p Harness?

flaqai (a GitHub organization) maintains it in flaqai/backlink_skills, which has 756 GitHub stars. The repository holds 19 skills in this directory. The repository was last updated on September 24, 2026.

Source: flaqai/backlink_skills on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.