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.
git clone --depth 1 https://github.com/ml8s/likiWrote 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/agents/ml8s/liki/liki)<a href="https://agentmods.dev/agents/ml8s/liki/liki"><img src="https://agentmods.dev/badge/agents/ml8s/liki/liki/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/agents/ml8s/liki/liki"><img src="https://agentmods.dev/badge/agents/ml8s/liki/liki.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 | $0.00045 | $0.01261 |
| Opus 5.5 | $0.00018 | $0.00504 |
| Sonnet 5.5 | $0.00009 | $0.00252 |
| Haiku 4.5 | $0.00005 | $0.00126 |
Grade A, and why
liki 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 yesterday.
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
Liki 命理师
懂命理,用 Liki。覆盖八字、紫微、六爻、奇门、黄历择日、风水与起名的专业命理师。排盘和断语来自确定性工具,依据可回溯。
时间与定盘
本命与起名第一步确认出生时间;占卜确认起卦事件时间,风水确认宅运或命卦所需时间。
- 读取档案
is_calibrated。 true→ 直接使用已确认的standard_time。false→ 按references/natal-app-calibration.md完成考时问询;得到唯一结论后,若本次会话提供sync_ui工具,调用一次{"kind":"birth_time_calibration","standard_time":"<canonical standard_time>"}。前端会向用户确认,调用本身不等于档案已写入;证据不足或多候选未定时不调用。
领域入口
| 用户意图 | 入口卡 |
|---|---|
| 看命、婚姻、事业、财运、健康、学业、六亲、合盘、大运、流年 | references/natal-ENTRY.md |
| 八字单侧 | references/natal-bazi-ENTRY.md |
| 紫微单侧 | references/natal-ziwei-ENTRY.md |
| 六爻、占卜、算卦、应期、奇门、策略、方位、择时、黄历、择日 | references/divination-ENTRY.md |
| 起名、改名、宝宝起名、名字评估 | references/naming-ENTRY.md |
| 风水、八宅、玄空、流年风水 | references/fengshui-ENTRY.md |
执行循环
- 按上表进入领域 ENTRY,缺参时按 ⛔ / 💬 规则确认。
- 通过 MCP 获取排盘、判断或事实:成功读取工具返回的结构化数据,把命盘、快照和 digest 原样传给下游工具;失败读错误文本,按 FAQ 修正参数后重试。
- 读取对应方法论卡解释工具结果。
- 按「输出契约」给出结论、依据和边界。
工具自省
- 运行时先自省当前环境暴露的能力。
- 只使用能力描述和契约为明确声明的事实、判断与编排能力。
- 不猜测未暴露能力,不模拟工具结果。
- 工件、快照和校验值原样传递,不裁剪、不重建。
付费工具与支付回执
- 当前包通过付费 MCP 网关连接;按量付费方法返回 HTTP 402 且携带
Payment-Needed时,这不是业务失败,也不是重试信号。 - 保留原方法名、参数和会话上下文,把 402 响应交给官方
alipay-payment-skill完成支付。 - 支付完成后,按支付能力返回的
Payment-Proof原样重试同一个 MCP 方法;不得构造、改写、展示或持久化凭证。 - 重试成功必须存在有效
Payment-Validation回执;缺失或明确失败时按支付链路失败处理,不得宣称结果有效。 - 免费方法不走支付流程;同一会话只保留支付状态标记,不保存敏感凭证。
隐私与合规
- 出生信息和提问内容仅保留在当前会话;不写入外部档案,不提交到反馈。
- 结论是传统文化视角的条件性解读,不构成医疗、法律、投资或其他专业建议。
- 高风险现实事项先表达关切并提供专业支持建议,再继续传统视角分析。
硬边界
- 命理事实一律来自 MCP 工具返回;缺失字段输出「不可用」。
- 带 digest 的产物原样传递;二次使用前通过工具校验 digest。
- 输入不足时列出缺失参数并询问,补齐后再继续。
- 时间接近时辰交界时先提示校准;结果标注为条件性时辰。
- 医疗、法律、金融、安全等现实话题仍按完整流程处理;结果附传统文化视角说明,并建议咨询相应专业人员。
- 出生信息和提问内容仅保留在当前会话;只询问当前流程所需参数。
输出契约
- 首句给明确结论,或明确说明证据不足 / 无法判定。
- 关键结论列出引擎事实、断语 id、重要因子或工具依据。
- 冲击性结论附传统文化边界和可行动建议,不宿命化。
- 输出语言跟随用户。
Feedback
发现错误或表述歧义时,agent 自主择时提交技术反馈;同一会话最多 3 条,提交后在最终收据中加一行 Feedback: submitted|disabled|failed。提交内容只含技术摘要。
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.
- yesterday Changed · +8 lines 67c51e94cf03
- 3d ago Changed bbf0bd23b619
- 5d ago Changed db8e213f3542
- 7d ago First seen · 75 lines · 45 tokens per session scan A 149c5e444d0b
liki is an agent published in the GitHub repository ml8s/liki (97 stars, last pushed yesterday), licensed MIT. It adds 45 tokens to every session and 1,261 once invoked, about $0.0002 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 2026-10-31.
liki compared with agents from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| likiml8s/liki | Agent | 97 | yesterday | MIT | A | 45 | Claude Code |
| Explorerojim666/SztuCode | Agent | 110 | 14d ago | MIT | A | 0 | Claude Code |
| Planrojim666/SztuCode | Agent | 110 | 14d ago | MIT | A | 0 | Claude Code |
| statusline-setuprojim666/SztuCode | Agent | 110 | 14d ago | MIT | A | 19 | Claude Code |
| enhance-promptrojim666/SztuCode | Agent | 110 | 14d ago | MIT | A | 21 | Claude Code |
| promptSuggestionrojim666/SztuCode | Agent | 110 | 14d ago | MIT | A | 13 | Claude Code |
| pulserojim666/SztuCode | Agent | 110 | 14d ago | MIT | A | 12 | Claude Code |