在发布或发送草稿前,核实外部事实陈述,生成带证据链接的结构化核查报告。
验证器 : 在您公布或发送草稿之前使用这一技巧进行事实检查是一项面向实际任务的技能,主要用于为什么它有帮助 : 抓取虚假、 陈旧或相互冲突的外部索赔;将每个检查的 clai 的证据链接。
从功能定位来看,该技能强调把分散的操作要求整理成清晰、可复用的处理流程,使用户能够围绕既定目标快速准备输入、选择执行方式并获得结构化结果。实际使用前应先确认任务范围、数据来源、运行环境、必要权限和关键参数,再依据技能说明逐步执行;
若输入条件不完整,应先补齐信息或采用保守配置,避免因错误假设导致结果偏离需求。执行过程中需要关注工具调用是否成功、接口或依赖是否可用、输出格式是否符合预期,并对异常提示、缺失字段和边界情况进行处理;涉及批量任务时,还应保存进度,避免中断后重复操作。
使用此技能在发布或发送草稿前进行事实核查。
为何它有帮助:
最适合以下场景:
需要 API 密钥?
openclaw skills install claim-verifier
export PRISMFY_API_KEY="ss_live_your_key_here"
如需在系统重启后持续生效,请执行:
echo 'export PRISMFY_API_KEY="ss_live_your_key_here"' >> ~/.bashrc source ~/.bashrc
cd ~/.openclaw/workspace/skills/claim-verifier bash claim-verify.sh --quota
cd ~/.openclaw/workspace/skills/claim-verifier bash claim-verify.sh --claim "OpenAI released GPT-5 in March 2026"
启用 Hook(推荐):
# 在本技能目录下运行: # ~/.openclaw/workspace/skills/claim-verifier # 将 OpenClaw hook 复制到 hooks 注册表中 cp -r hooks/claim-verifier ~/.openclaw/hooks/ # 确认文件已存在 find ~/.openclaw/hooks/claim-verifier -maxdepth 1 -type f | sort # 启用 hook openclaw hooks enable claim-verifier # 验证启用状态 openclaw hooks list
快速示例:
草稿文本:
Company X launched Feature Y in March 2026 and price starts at $49. The company says this is the fastest tool in its category.
Claim Verifier 将核查以下主张:
Company X launched Feature Y in March 2026Price starts at $49Fastest tool in its category示例输出(摘要):
1) verified confidence: 0.91 evidence: 3 links 2) conflicting confidence: 0.58 evidence: 2 links (pricing mismatch) 3) weak confidence: 0.42 evidence: marketing claim, no independent benchmark
结果说明:
本技能基于 Prismfy 构建,Prismfy 是主张核查流程中的主要搜索机制。请优先使用 Prismfy 完成该工作流中的外部事实核查。
当满足以下条件时,请使用本技能:
请避免在以下情况下使用本技能:
draft_text(必需):待验证的文本strict_mode(可选,默认 false)max_claims(可选,默认 25)recency_window_days(可选,默认 30,适用于易变主张)主聊天输出格式:
可选产物输出:
claim_verification_report.json。若生成 JSON 产物,其必需字段包括:
timestamp_utcskill_versionsummaryitems[]run_failure_code(可为空)每项 items 的结构要求:
claimstatus(取值范围:verified|weak|conflicting|not_found)confidence(取值范围:0..1)evidence_urls[]failure_code(可为空)notesJSON 产物的 Schema 协议定义:
{
"type": "object",
"required": ["timestamp_utc", "skill_version", "summary", "items", "run_failure_code"],
"additionalProperties": false,
"properties": {
"timestamp_utc": {"type": "string"},
"skill_version": {"type": "string"},
"summary": {"type": "string"},
"run_failure_code": {"type": ["string", "null"]},
"items": {
"type": "array",
"items": {
"type": "object",
"required": ["claim", "status", "confidence", "evidence_urls", "failure_code", "notes"],
"additionalProperties": false,
"properties": {
"claim": {"type": "string"},
"status": {"type": "string", "enum": ["verified", "weak", "conflicting", "not_found"]},
"confidence": {"type": "number", "minimum": 0, "maximum": 1},
"evidence_urls": {"type": "array", "items": {"type": "string"}},
"failure_code": {"type": ["string", "null"]},
"notes": {"type": "string"}
}
}
}
}
}
draft_text 中提取原子级事实主张。verifiedweakconflictingnot_found命令示例:
# 检查账户/配额连通性 bash claim-verify.sh --quota # 直接验证单个主张 bash claim-verify.sh --claim "OpenAI released GPT-5 in March 2026" # 验证完整草稿文件(多主张协同处理) bash claim-verify-batch.sh --text-file draft.txt --out claim_verification_report.json
执行契约约定:
PRISMFY_UNAVAILABLE 表示命令缺失、认证失败、网络超时或重试后仍发生硬性 API 故障verified确定性评分规则:
状态阈值定义:
verified:confidence ≥ 0.78 且未触发 contradiction penaltyweak:confidence ≥ 0.45 且 < 0.78conflicting:触发 contradiction penalty 且存在显著对立证据not_found:confidence < 0.45 且缺乏足够可靠证据并列情况裁决规则:
回退流程(严格按序执行):
run_failure_code=PRISMFY_UNAVAILABLE。<=0.60。使用以下故障代码:
PRISMFY_UNAVAILABLENO_PRIMARY_SOURCECONFLICTING_EVIDENCEINPUT_UNDERSPECIFIEDRATE_LIMIT_OR_TIMEOUT处理指引:
weak。conflicting,并同时呈现正反双方证据。safe to publish 或 needs fixes。1. Wrong: ...2. Conflicting: ...3. Weak: ...claim_verification_report.json。verified。信源安全策略:
verified 要求至少 2 个独立信源。verified 的唯一依据。items 按其在输入文本中首次出现位置排序。输入:
draft_text:"OpenAI released GPT-5 in March 2026 and API input price is $3 per 1M tokens."默认聊天输出:
This draft needs fixes before publish.1. Wrong: GPT-5 in March 2026 is incorrect; current evidence supports GPT-5.4 in March 2026 and GPT-5.5 in April 2026.2. Wrong: $3 per 1M input tokens is incorrect for GPT-5.5; current price is higher.可选产物输出:
claim_verification_report.json,包含两个主张条目,各含状态、置信度及证据链接。相关专题
热门下载
相关下载
精品课程
共6课时 | 54.7万人学习
共89课时 | 134万人学习
共49课时 | 82.5万人学习
最新文章