Clowder AI is a self-hosted workspace where AI agents from different model families work together as a persistent team, retaining identities, shared evidence, and memory across tasks. It is for people who want to coordinate multiple AI agents without repeatedly rebuilding their context.
Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/zts212653/clowder-ainpx agentmods add skills/zts212653/clowder-ai/merge-gateWrote 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/zts212653/clowder-ai/merge-gate)<a href="https://agentmods.dev/skills/zts212653/clowder-ai/merge-gate"><img src="https://agentmods.dev/badge/skills/zts212653/clowder-ai/merge-gate/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/zts212653/clowder-ai/merge-gate"><img src="https://agentmods.dev/badge/skills/zts212653/clowder-ai/merge-gate.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 2 findings, up to high
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- high Tool Misuse · line 83 Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).Fix: Validate all tool parameters against an allowlist. Reject dangerous parameter values (shell=True, --force, -rf /) and use safe defaults.
- high Tool Misuse · line 592 Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).Fix: Validate all tool parameters against an allowlist. Reject dangerous parameter values (shell=True, --force, -rf /) and use safe defaults.
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 | $0.00046 | $0.15372 |
| Opus 5 | $0.00023 | $0.07686 |
| Sonnet 5 | $0.00009 | $0.03074 |
| Haiku 4.5 | $0.00005 | $0.01537 |
Grade C, and why
merge-gate scanned grade C with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
rm -rf "$sandbox" How it starts
The opening of the file, as written. The whole thing — 708 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Merge Gate
SOP definition:
sop-definitions/development.yamlstagemerge。
合入 main 的流程：先锁定风险档、验证命令与独立 review source，再执行对应门禁。PR 是载体，不是自动触发 local + cloud + guardian 三连的理由。
Lane 0：Co-Creation Docs PR
先检查是否已有成功的 pnpm classify:co-creation-docs 证据，且输出同时满足：
lane=co_creation_docsdelivery=pull_request- changed files 与 PR diff 完全一致
满足时，直接消费 classifier 的 validation / cloudReview / fullGate 结论：
- 跑 classifier 返回的全部
validation命令 +git diff --check。 - 新实质内容需要非作者内容 review；已有内容 verdict 或可证明机械合并用 continuityProof 复用，不为 SHA 字符串变化重开 reviewer。
cloudReview=required才触发 cloud；skip时在 PR 留 classifier 原因。家规 / SOP / skill 纯文字通常由有状态 local reviewer 覆盖治理语义，不把“免 cloud”当需要申请的特权。fullGate=required才跑pnpm gate；skip时 evidence manifest 记录 docs validation 命令。- evidence 闭合后由在场 merge owner 执行
gh pr merge --squash；不再额外召唤一只猫只为按 merge 按钮。无 F 号时跳过 Feature Doc Truth post-merge sync。
任一 changed file 不匹配、classifier 缺失/失败或输出 lane=regular_development → 退出本节，走下方风险路由。行数不能作为 Lane 0 证据。
核心知识
Risk-Routed Merge 门禁 5 条（全部满足才能合入）
- PR body 写清五轴风险判断：行为面 / 数据 / 安全 / 契约 / 不可逆；默认最小安全动作，升档理由可查。
- 至少一个非作者独立 review source（local 或 cloud）有明确 verdict；仅在不同高风险面需要不同视角时叠加，愿景守护另按 feature-close 触发。
- 所有 P1/P2 已修复，并由提出 finding 的活跃 source 覆盖当前 HEAD（含 Harness Diet Rebase Continuity 的 continuityProof 桥接）。
- 适用的 feature / BACKLOG 真相源没有过度声称，PR 载体与 changed files 匹配。
- 与风险匹配的 gate 全绿：低 / 中风险用 targeted commands；安全、鉴权、生产数据、迁移、外部契约或不可逆风险用
pnpm gate全量。
默认只选一个合适的独立 source：家里语境与治理语义优先 local；context-blind 安全 / 契约代码扫描优先 cloud。动作类型（“开了 PR”“改了代码”）不是叠加理由。
Review Continuity Guard（review 是否真的覆盖当前 HEAD）
pnpm gate、rebase、fixup、biome 格式化刷新等都可能让 HEAD 变化。HEAD 变化只触发 provenance 判定，不自动等于 re-review：先分清 review 后是否真的改了本 PR 的内容、base 前进是否与本 PR 有逻辑关联；两者都没有，或只有可机械证明的派生物重建 / 规范化，旧 review 用 continuityProof 桥接。只有真实的作者 delta 或相关 base delta 回 active source，而且只看那一小块。
昂贵 gate 连续性（ROI 硬边界）：一次 full gate 绑定“作者 patch + gate 开始时冻结的 base”，而不是绑定会继续移动的 origin/main 字符串。full gate 已完整通过后，若后续只是纯 rebase、作者 patch 不变或 patch-equivalent，且 C2 证明 base 增量无关联，则 rebase 后只跑风险匹配的 targeted continuity checks；禁止仅因 main 又前进而重跑 full gate。上一轮 full gate 未完整通过只会使旧 receipt 不可复用，不会单独把本来属于 targeted 的改动升级为 full。pnpm gate 会在冻结 base 后从 Git diff、既有 terminal/stage receipt 与失败输出尾部自动选择车道：只有作者实质 delta、相关或无法判定的 base delta、冲突中的语义取舍、相关旧失败，或真实高风险面才重新运行 full gate。机器判为 targeted 时，命令会在申请 full-gate 资源前退出；必须另行刷新受影响检查与全仓跨包 typecheck，并把命令写进 evidence manifest。classifier 只选车道，不代替这些绿色证据。
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.
- 5d ago Changed · -34 lines 07c5a3fe3638
- 6d ago Changed · +3 lines 2385abe37fc5
- 10d ago First seen · 739 lines · 46 tokens per session scan C 540ca0325178
merge-gate is a skill published in the GitHub repository zts212653/clowder-ai (2,956 stars, last pushed today), licensed MIT. It adds 46 tokens to every session and 15,372 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it C with 1 finding (recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
adversarial-reviewer
Adversarial code review that assumes bugs exist and hunts for them. Use when asked to review code, find bugs, audit for correctness, stress-test a PR, or when someone says "tear this apart" or "what's wrong with this". Give no benefit of the doubt — every line is guilty until proven innocent.
go-testing
Trigger: Go tests, go test coverage, Bubbletea teatest, golden files. Apply focused Go testing patterns.
semgrep-rule-variant-creator
Creates language variants of existing Semgrep rules. Use when porting a Semgrep rule to specified target languages. Takes an existing rule and target languages as input, produces independent rule+test directories for each language.
brooks-sweep
Full-sweep mode: runs a unified analysis across all quality dimensions — code decay, architecture, tech debt, and test quality — then applies fixes directly to the codebase. Safe changes are auto-applied; risky changes are confirmed before execution. Drawing on twelve classic engineering books. Triggers when: user…
include-test-files-that-assert-on-behavior-being-changed-in-decl
When delegating a task affected by this skill, include.
plugin-review
Review plugin quality with tiered checks and dependency scoping. Use for PR and pre-release audits.