a-share-dragon-tiger-report@2.0.1
CAUTION· static onlyPublisher claims: "生成 A 股「涨停板龙虎榜深度研判报告」提示词 v4.0「宏观锚定 · 五维共振 · 水印隔离 · GCCEO 全量整合 · 投行级量化研判」 (2026-11-12 定稿)。口径 v2.3 六条永久冻结 + KNOWLEDGE 45 键;三层只读旁路叠加: v3.0 五维共振(L0宏观16条只下调 / 策略因子13条 / X-主题17方向三层白名单 / 短线金手24条 / 三类水印,自检14/14) + v3.1 GCCEO 4.0 全量整合(十域91技能 / 投行A1-A10 / 出海G1-G10 / AI超能力8 / 方法论16 / 五步决策框架,自检23/23) + v4.0 投行级量化研判(真值算+缺口标注:估值锚/利润质量/五维形态/Pre-Mortem/缺口台账/全链路,自检25/25)。 GCCEO 以「研判视角映射」接入:只输出标签/清单/风险提问/叙事脚手架,无剔除权、无打分权、无仓位上调权; v4.0 量化层真值算、缺口如实标注,**绝不编造 DCF/LBO/ROE/净现比**,不生成 BUY。 四层漏斗 L1→L2→L2.5→L3(半年口径 + 增速红线 + 利润质量红旗≥2 强否决)一字不改; 情绪五阶段仓位 30/60/70/20/0% 不变。 BII 双栏默认 20 页(前14页 v2.3 逐字不改 + 第15/16页 v3 附录 + 第17/18页 GCCEO 附录 + 第19/20页 v4.0 投行量化附录); 四态回滚:gceo_quant.enabled=false→18页、gceo.enabled=false→16页、v3.enabled=false→14页(单向强依赖 gceo_quant⊂gceo⊂v3)。 输出 HTML / PDF;样式 A(classic,含 PPTX)仅在明确要求时补出。 内置「十五五 18 方向(102板块+12SW1+70股)」与 Anthropic S-1 → A股 AI 链映射; 蒸馏语料 31 份 + 4 本经典 + 5 份宏观/趋势研报;GCCEO 源 10 域 91 技能。仅交易日生成,非交易日直接说明休市并结束。 v3.1.1 编排层:固化「当日主研判 / 次日预判核验 / 投资大师圆桌」三条流水线的触发前提、 数据契约、产物路径与权限边界(第十八章)。 v3.1.2 资金规模层:按本金规模自适应换结构(小本金 / 机构级 ≥1亿),金额三档单位、 图表刻度、第 10 页由「α 预算纪律」重构为「资金容量与执行约束」(真实 20 日均成交额实算容量表); 判定内核与冻结口径零改动,自检与页数不变。 v3.2 选股源主口径:selection_mode=limitup 时从当日涨停梯队 + 龙虎榜净买 + 主线资金找, 超跌池降级为位置对照;旧漏斗与冻结阈值一字不改,oversold 可完整回滚。 v3.2.1 涨停源深化:validate() 数据契约六项校验 / verify_covered() 真跑 L2.5·L3 冻结闸门 / capacity_check() 联动 §18.7 资金容量层;第 13 页整表改涨停源观察榜(动作恒 WATCH), 第 4 页声明 TOP8 截断与闸门实跑,第 13 页说明条登记契约与容量结论;自检 11/11。 v4.0.3 跨会话长期实盘口径:本金 ¥100,000(取代旧 7,000 元),周目标 ¥15,000、 单笔最大亏损 ¥5,000;比例纪律与两套闸门不变,7 亿机构容量口径独立不串味(§18.11)。 v4.0.4 统一深度投研报告流水线:四阶段 A/B/C/D 合并为「一次触发 · 一份主交付」, 2026-11-13 实跑基线(64 内联 SVG / 31 表 / 57 页 A4)、三链 5/20 日净额口径修正 与三重校验清单见 §18.12 及 docs/unified-deep-research-pipeline-20261113.md。 v4.0.5 安全加固:新增 §18.13 工具权限边界与安全声明(行情只读 / 本地文件只写自身产物 / 无凭证访问 / 无自动远端发布);安装脚本移除 `rsync --delete` 改为备份后逐文件覆盖; 示例 JSON 与样本报告明确标注为 2026-11-13 快照,禁止沿用为当期真值。 **不构成投资建议**。" Observed in sandbox: static analysis only — behavior not yet observed.
Coverage
Static only; nothing was run yet.The label rests on the code, the instruction files and the publisher. The honeypot run replaces this verdict when it completes.
Is [email protected] safe to install?
KillerSkills analysed it statically (a sandbox run is pending) on 2026-11-07 and scored it 69/100 (Caution). It is not known to be malicious, but it has behaviour or code a security policy should decide on before install. Summary: static analysis only — behavior not yet observed.
Behavior trace
Queued for honeypot run · static analysis onlyThis version has not been detonated yet. The score above counts behavior as unknown (50/100), so it cannot read Verified until the honeypot run completes and replaces this verdict.
Static findings
res = subprocess.run(cmd, cwd=str(TEMPLATES), env=env, capture_output=True, text=True)Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.No other findings across 13,092 lines in 39 files.
Cost to the agent
Tokens estimated at four characters each. Skills load their instruction text when triggered; the figure is what that costs.
Known vulnerabilities
Vulnerability advisories (OSV) cover npm and PyPI packages; this registry has no advisory database.
What would make this Verified
11 points to 80 · weights: behavior 45%, static 25%, prompt 15%, publisher 15%- Unclaimed publisher≈ +8 on the totalClaim yjkj999999: a verified publisher gets +50 on publisher trust. It is free and usually instant with GitHub sign-in.
- Python subprocess execution detected≈ +2 on the totalBuild commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.