[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6MGQJPCXENT1QPPP73NVMBD":3,"skill-same:01M6MGQJPCXENT1QPPP73NVMBD":90},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":23,"publishedAt":46,"linkOnly":47,"markdownContent":48,"markdownBody":49,"frontMatter":50,"refName":60,"commitSha":61,"skillDirectory":62,"forksCount":63,"repositoryHtmlUrl":64,"repositoryHomepageUrl":65,"ownerHtmlUrl":66,"ownerBlogUrl":67,"ownerDescription":68,"updatedAt":69,"install":70,"resources":83},"01M6MGQJPCXENT1QPPP73NVMBD","unified-ai-gateway",null,"Operate and evaluate Unified AI System through fifteen governed MCP tools, including provider-free prompt enhancement, while preserving fake-provider, authorization, and evidence boundaries.","hashgraph-online\u002Fawesome-codex-plugins","hashgraph-online","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F160371962?v=4","01M6CNB6NRJVNNGN14TR01Q6D3","HOL (Hashgraph Online)","01M6JT1HSKKG9V68KQAQGBA7XP",1566,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fhappy520ai\u002Funified-ai-system\u002Fskills\u002Funified-ai-gateway\u002FSKILL.md",[18],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",[24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45],"gateway","ai gateway","unified ai gateway","master","2026","enhancement","tree","operate","ai-gateway","free","critical","self","provider","MCP","self-hosted","evidence","prompt","system","Tools","license","ml","official","2026-11-10 10:12:30",false,"---\nname: unified-ai-gateway\ndescription: Operate and evaluate Unified AI System through fifteen governed MCP tools, including provider-free prompt enhancement, while preserving fake-provider, authorization, and evidence boundaries.\nlicense: Apache-2.0\nmetadata:\n  author: happy520ai\n  category: ai-ml\n  date_added: \"2026-08-01\"\n  license_source: https:\u002F\u002Fgithub.com\u002Fhappy520ai\u002Funified-ai-system\u002Fblob\u002Fmaster\u002FLICENSE\n  risk: critical\n  source: https:\u002F\u002Fgithub.com\u002Fhappy520ai\u002Funified-ai-system\u002Ftree\u002Fmaster\u002Fskills\u002Funified-ai-gateway\n  source_repo: happy520ai\u002Funified-ai-system\n  source_type: official\n  tags: \"ai-gateway, codex, mcp, self-hosted, governance\"\n  tools: codex\n---\n\n# Unified AI Gateway\n\n## Overview\n\nUse the official `unified-ai-system` MCP server to inspect and exercise a local\nAI gateway without provider credentials. This skill file provides operating\nguidance; it does not install the server or change Codex configuration by\nitself. The official Codex plugin bundles the MCP definition, while skill-only\ninstallations require the manual setup below.\n\n## Version Note\n\nThese are two different things and they are not equal today:\n\n- **Current release: `v0.8.0`.** It declares and ships fifteen tool names, and the\n  [60-second demo command](https:\u002F\u002Fgithub.com\u002Fhappy520ai\u002Funified-ai-system#try-it-in-60-seconds)\n  in the README names that version. Read it live with\n  `node tools\u002Fverify-image-roster.mjs 0.8.0`, which reports the roster from the\n  image bytes rather than from this file. That command needs a clone of the\n  repository, run from its root: the script ships with the source, not inside an\n  installed skill bundle. Without a clone, the same roster is published at\n  https:\u002F\u002Fhappy520ai.github.io\u002Funified-ai-system\u002Fverify-mcp-docker-image.html.\n- **Newest content review: `0.8.0`** ([review](https:\u002F\u002Fgithub.com\u002Fhappy520ai\u002Funified-ai-system\u002Fblob\u002Fmaster\u002Fdocs\u002Fsecurity\u002Fmcp-image-review-0.8.0.md)),\n  read from the published layer tarballs with `node tools\u002Finspect-image-filesystem.mjs 0.8.0` - no Docker\n  engine and nothing executed. It found one thing worth knowing before you pull: the `linux\u002Farm64` tag carries\n  x86-64 native modules, including `better_sqlite3.node`, so request `--platform linux\u002Famd64`.\n- **Reviewed and pinned below: `0.4.9`.** The inspection procedure in this file\n  pins that image's recorded digests because `0.4.9` is the newest version reviewed by that procedure, which\n  exports a container filesystem with Docker. The 0.8.0 review above uses a different method and pins its own\n  digests, so the table below still describes 0.4.9 until it is rewritten against a 0.8.0 export.\n  `0.4.9` carries 9 of the fifteen names: the model-backed enhancement, knowledge\n  retrieval and workflow execution tools arrived at 0.5.0, and the three\n  governance tools at 0.8.0.\n\nDo not substitute a mutable tag for a pinned digest, and do not move the pin to a\nnewer version just because this file looks out of date: a new content review is\nrequired first, and the pinned identity is only as good as the review that backs it.\n\n## Prerequisites And Setup\n\n1. Confirm that Codex CLI and Docker are installed and Docker is running.\n2. If the nine tools are already visible, skip setup and do not register a\n   duplicate server.\n3. Explain the first stage: it downloads one reviewed platform from the\n   immutable `0.4.9` multi-platform index into Docker's cache, inspects its\n   metadata and layer history, creates but never starts a temporary container,\n   exports its root filesystem, removes that temporary container, and writes an\n   inspection inventory to a temporary directory. The reviewed platforms are\n   linux\u002Famd64 and linux\u002Farm64. Obtain explicit user approval for those download\n   and inspection changes only.\n4. After that first approval, pull the reviewed platform manifest and complete\n   the inspection. Do not execute the image or register it yet:\n\n```bash\nIMAGE='ghcr.io\u002Fhappy520ai\u002Funified-ai-system\u002Fmcp-server@sha256:751a0d32acd2d6b1da6ad9ac67987fbd1ff36ce26b7160014d8605f18b7907b3'\nPLATFORM='linux\u002Famd64' # Use linux\u002Farm64 only on a reviewed ARM64 engine.\nREVIEW_DIR=\"$(mktemp -d)\"\n\ndocker pull --platform \"$PLATFORM\" \"$IMAGE\"\ndocker image inspect \"$IMAGE\" --format 'Id={{.Id}} OS={{.Os}} Architecture={{.Architecture}} User={{json .Config.User}} Entrypoint={{json .Config.Entrypoint}} Cmd={{json .Config.Cmd}} Labels={{json .Config.Labels}}'\ndocker image history --no-trunc \"$IMAGE\" > \"$REVIEW_DIR\u002Fimage-history.txt\"\n\nREVIEW_CONTAINER=\"$(docker create --platform \"$PLATFORM\" --pull never --entrypoint \u002Fbin\u002Ftrue \"$IMAGE\")\"\ndocker export --output \"$REVIEW_DIR\u002Frootfs.tar\" \"$REVIEW_CONTAINER\"\ndocker rm \"$REVIEW_CONTAINER\"\n\ntar -tf \"$REVIEW_DIR\u002Frootfs.tar\" > \"$REVIEW_DIR\u002Frootfs-files.txt\"\nmkdir -p \"$REVIEW_DIR\u002Frootfs\"\ntar --same-permissions -xf \"$REVIEW_DIR\u002Frootfs.tar\" -C \"$REVIEW_DIR\u002Frootfs\"\nfind \"$REVIEW_DIR\u002Frootfs\u002Fapp\" -type f -print > \"$REVIEW_DIR\u002Fapp-files.txt\"\n: > \"$REVIEW_DIR\u002Fapp-links.txt\"\nwhile IFS= read -r -d '' APP_LINK; do\n  ls -ld -- \"$APP_LINK\" >> \"$REVIEW_DIR\u002Fapp-links.txt\"\ndone \u003C \u003C(find \"$REVIEW_DIR\u002Frootfs\u002Fapp\" \\( -type l -o -type f -links +1 \\) -print0)\n: > \"$REVIEW_DIR\u002Fnative-binaries.sha256\"\nwhile IFS= read -r -d '' NATIVE_BINARY; do\n  sha256sum -- \"$NATIVE_BINARY\" >> \"$REVIEW_DIR\u002Fnative-binaries.sha256\"\ndone \u003C \u003C(find \"$REVIEW_DIR\u002Frootfs\u002Fapp\" -type f -name '*.node' -print0)\nfind \"$REVIEW_DIR\u002Frootfs\" -type f \\( -perm -0100 -o -perm -0010 -o -perm -0001 \\) -print > \"$REVIEW_DIR\u002Fexecutable-files.txt\"\nfind \"$REVIEW_DIR\u002Frootfs\" -type f \\( -perm -4000 -o -perm -2000 \\) -print > \"$REVIEW_DIR\u002Fsuid-sgid-files.txt\"\nfind \"$REVIEW_DIR\u002Frootfs\u002Fapp\" -type f \\( -name '.env' -o -name '.env.*' -o -name '*.pem' -o -name '*.key' -o -name '*.p12' -o -name '*.pfx' -o -path '*\u002F.ssh\u002Fid_*' \\) -print > \"$REVIEW_DIR\u002Fcredential-like-files.txt\"\ngrep -RInHE --include='package.json' '\"(preinstall|install|postinstall|prepare|prepack|postpack)\"' \"$REVIEW_DIR\u002Frootfs\u002Fapp\" > \"$REVIEW_DIR\u002Flifecycle-hooks.txt\"\ngrep -RInE 'child_process|spawn\\(|fetch\\(|AI_GATEWAY_MCP_URL|process\\.env|writeFile|appendFile|unlink|rm\\(' \"$REVIEW_DIR\u002Frootfs\u002Fapp\u002Fpackages\u002Fmcp-server\u002Fsrc\" \"$REVIEW_DIR\u002Frootfs\u002Fapp\u002Fpackages\u002Fshared-sdk\u002Fsrc\" > \"$REVIEW_DIR\u002Fruntime-sensitive-code.txt\"\n```\n\nIf `sha256sum` is unavailable, use the platform's SHA-256 utility and preserve\nthe same report. Keep the review directory until the report is accepted; its\ndeletion is another filesystem change and requires approval for the exact path.\n\n5. Read every generated inventory and report the inspection before proceeding.\n   Compare it with the versioned\n   [image content review](https:\u002F\u002Fgithub.com\u002Fhappy520ai\u002Funified-ai-system\u002Fblob\u002Fmaster\u002Fdocs\u002Fsecurity\u002Fmcp-image-review-0.4.9.md).\n   Require OCI index digest\n   `sha256:751a0d32acd2d6b1da6ad9ac67987fbd1ff36ce26b7160014d8605f18b7907b3`.\n   For linux\u002Famd64, require manifest digest\n   `sha256:ff6cf988b01d5fb2e97aabe8e952f6a303dcffe650df5b4dcb0ba3d51ee88c06`\n   and config digest\n   `sha256:0c2c0c7b9c7fb7ca24c73d9a903bcf719b079a0b285a3a3269ee3ae059905e97`.\n   For linux\u002Farm64, require manifest digest\n   `sha256:90318b9e373820f863c1c1addc759be4b5ce186f2ecb6232ee502fad7c6613de`\n   and config digest\n   `sha256:c2047eb63fdc42bcb16d53fca17d78a4a6fb355cf6320b9aa6688e594371054f`.\n   Require source `https:\u002F\u002Fgithub.com\u002Fhappy520ai\u002Funified-ai-system`, revision\n   `342a47313927870bcc696be13c9e5fb922062dac`, version `0.4.9`, license\n   `Apache-2.0`, entrypoint `docker-entrypoint.sh`, and command\n   `node packages\u002Fmcp-server\u002Fsrc\u002Findex.js`.\n\n   Report these reviewed risks explicitly: the image uses the default root\n   user; includes Debian shell\u002Fpackage utilities and 11 base-image SUID\u002FSGID\n   files; contains 522 internal pnpm links, three native Node binaries, and eight\n   lifecycle-hook declarations; and starts a child gateway with loopback HTTP.\n   The optional `AI_GATEWAY_MCP_URL` can make an HTTP or HTTPS connection only\n   when explicitly passed. The registered command below passes no host files,\n   environment variables, or ports and disables container networking. Stop on\n   any mismatch, unexpected link, credential-like file, native binary, hook,\n   privileged file, or sensitive-code behavior.\n6. Explain the second stage: it persists a Codex MCP configuration and permits\n   Codex to launch the inspected image in a later task. Obtain a separate\n   explicit approval for registration and activation; the download approval\n   does not carry over.\n7. After that second approval, register the reviewed platform digest with\n   pulling, container networking, Linux capabilities, and privilege escalation\n   disabled, then inspect the stored configuration:\n\n```bash\nIMAGE='ghcr.io\u002Fhappy520ai\u002Funified-ai-system\u002Fmcp-server@sha256:751a0d32acd2d6b1da6ad9ac67987fbd1ff36ce26b7160014d8605f18b7907b3'\nPLATFORM='linux\u002Famd64' # Match the reviewed platform inspected above.\ncodex mcp add unified-ai-system -- docker run --rm -i --pull never --platform \"$PLATFORM\" --network none --cap-drop ALL --security-opt no-new-privileges \"$IMAGE\"\ncodex mcp get unified-ai-system --json\n```\n\n8. Restart Codex or open a new task, then use `\u002Fmcp verbose` to confirm that all\n   nine tools are available - the pinned `0.4.9` image ships nine of the fifteen\n   names the current release declares. Remove the registration when it is no longer\n   wanted:\n\n```bash\ncodex mcp remove unified-ai-system\n```\n\nRemoving the registration does not remove the pulled image from Docker's\ncache. Treat image-cache deletion as a separate host-state change and obtain\napproval before doing it.\n\n## When to Use This Skill\n\n- Use when a user asks whether Unified AI System is healthy or ready.\n- Use when a user wants a credential-free gateway chat proof.\n- Use when a user asks about the gateway's knowledge, workflow, or workforce\n  surfaces.\n- Use when a user wants evidence from the bundled MCP tools rather than a claim\n  inferred from documentation or process exit codes.\n\nDo not use this skill for generic model comparisons, unrelated MCP servers, or\ndeploying a production gateway.\n\n## Workflow\n\n1. Confirm that the `unified-ai-system` MCP tools are available in the current\n   task. If they are absent, follow the approved setup above and wait for a\n   restarted or new task.\n2. Call `gateway_health`, then `gateway_readiness`, before attempting chat.\n3. Select the narrowest additional tool that answers the request.\n4. Report returned provider, execution mode, readiness, and blockers exactly.\n5. Separate transport success from product, production-readiness, autonomy, or\n   AGI claims.\n\n## Tool Map\n\nNine of these ship in the reviewed `0.4.9` image below; the six marked 0.5.0 and\n0.8.0 are in the current release and are absent from that older image.\n\nStatus and boundaries:\n\n- `gateway_health`: gateway health, provider mode, and the real-provider safety flag\n- `gateway_readiness`: first-run readiness for chat and the local gateway runtime\n- `knowledge_readiness`: knowledge infrastructure without loading or changing data\n- `workflow_health`: the governed workflow subsystem without starting a workflow\n- `workflow_actions`: workflow action definitions without invoking any action\n- `workforce_health`: the workforce subsystem without planning or executing work\n- `workforce_agents`: configured workforce agent descriptors without dispatching them\n\nDoing work locally, still with no provider call:\n\n- `gateway_prompt_enhance`: structures a plain-language request into a prompt\n  locally, without provider credentials or provider calls\n- `gateway_prompt_enhance_llm` *(0.5.0)*: semantic rewriting through a provider\n  when one is configured, falling back to the deterministic local engine when none is\n- `knowledge_retrieve` *(0.5.0)*: keyword search over the local knowledge base,\n  returning ranked chunks with citations; calls no provider\n- `workflow_run` *(0.5.0)*: the 3-step local workflow - retrieve knowledge, compose\n  a Markdown report, write a controlled artifact; calls no provider\n- `gateway_chat`: one chat request, accepted only when the gateway proves real\n  providers are disabled\n\nThe governed Agent surface, read-only *(0.8.0)*:\n\n- `agent_governance_status`: governance status through the authenticated Gateway\n  identity; no tenant override is accepted and it fails closed when that identity is\n  not authorized for platform status\n- `agent_governance_list`: only the governed Agents visible to that tenant, with no\n  tenant or owner override argument\n- `agent_governance_describe` (takes `agentId`): one Agent as seen by that tenant;\n  cross-tenant and missing identifiers stay indistinguishable on purpose\n\nCreating, executing, revoking, approving or activating anything in that surface\nremains a human REST\u002FSDK\u002FCLI operation and is not exposed to the model at all.\n\n## Example\n\n```text\nUser: Check whether the local gateway is ready, then prove chat works safely.\n\nAgent:\n1. Call gateway_health.\n2. Call gateway_readiness.\n3. Call gateway_chat only if both results prove fake-provider mode.\n4. Report provider, model, execution mode, response, and every blocker.\n```\n\n## Safety Boundaries\n\n- Keep the credential-free local fake provider as the default.\n- Never request, read, or transmit provider credentials through this skill.\n- Do not enable or call a real provider without explicit scoped authorization.\n- Treat MCP registration, image pulls, container creation, networking, and\n  teardown as host-state changes that require informed user approval.\n- Never substitute a mutable tag, a different OCI index, or an unreviewed\n  platform manifest for the reviewed `0.4.9` identities. Keep download and\n  inspection approval separate from registration and activation approval.\n- Keep `--pull never` in the registered command. If the reviewed image is\n  absent from the local cache, fail closed and return to the first approval\n  stage.\n- Keep `--network none`, `--cap-drop ALL`, and\n  `--security-opt no-new-privileges` in the registered command.\n- Do not claim production readiness, L5 autonomy, or AGI from a healthy handshake.\n- Treat a zero exit code as transport evidence, not proof that readiness gates\n  passed.\n\n## Limitations\n\n- This skill file does not bundle the MCP server, Docker image, or Codex\n  configuration. It only operates tools supplied by the separately installed\n  official integration.\n- It does not deploy, benchmark, or certify the gateway for production use.\n- The credential-free chat tool proves only the deterministic local fake path.\n- It does not configure real providers or handle provider credentials.\n- The published MCP image requires Docker.\n- The reviewed `0.4.9` path covers linux\u002Famd64 and linux\u002Farm64. Do not activate\n  another platform image without a separate content review.\n- The image runs as the container's default root user and bundles the gateway\n  source, package-manager tooling, native dependencies, and base-image\n  SUID\u002FSGID files. The registered command drops capabilities, prevents new\n  privileges, disables networking, and leaves the image in Docker's cache.\n- Existing Codex tasks may not hot-load a newly installed MCP configuration.\n\n## Troubleshooting\n\n- If the tools are missing after approved registration, inspect\n  `codex mcp get unified-ai-system --json`, then restart Codex or start a new\n  task.\n- If readiness is blocked, report the returned blocker instead of retrying chat\n  blindly.\n- If the runtime might use a real provider, stop before chat and keep the\n  session read-only.\n\n## Additional Resources\n\n- [Unified AI System](https:\u002F\u002Fgithub.com\u002Fhappy520ai\u002Funified-ai-system)\n- [60-second Codex MCP quickstart](https:\u002F\u002Fgithub.com\u002Fhappy520ai\u002Funified-ai-system\u002Fblob\u002Fmaster\u002Fdocs\u002Fcodex-mcp-quickstart.md)\n- [MCP server guide](https:\u002F\u002Fgithub.com\u002Fhappy520ai\u002Funified-ai-system\u002Fblob\u002Fmaster\u002Fpackages\u002Fmcp-server\u002FREADME.md)\n- [MCP image content review](https:\u002F\u002Fgithub.com\u002Fhappy520ai\u002Funified-ai-system\u002Fblob\u002Fmaster\u002Fdocs\u002Fsecurity\u002Fmcp-image-review-0.4.9.md)\n","# Unified AI Gateway\n\n## Overview\n\nUse the official `unified-ai-system` MCP server to inspect and exercise a local\nAI gateway without provider credentials. This skill file provides operating\nguidance; it does not install the server or change Codex configuration by\nitself. The official Codex plugin bundles the MCP definition, while skill-only\ninstallations require the manual setup below.\n\n## Version Note\n\nThese are two different things and they are not equal today:\n\n- **Current release: `v0.8.0`.** It declares and ships fifteen tool names, and the\n  [60-second demo command](https:\u002F\u002Fgithub.com\u002Fhappy520ai\u002Funified-ai-system#try-it-in-60-seconds)\n  in the README names that version. Read it live with\n  `node tools\u002Fverify-image-roster.mjs 0.8.0`, which reports the roster from the\n  image bytes rather than from this file. That command needs a clone of the\n  repository, run from its root: the script ships with the source, not inside an\n  installed skill bundle. Without a clone, the same roster is published at\n  https:\u002F\u002Fhappy520ai.github.io\u002Funified-ai-system\u002Fverify-mcp-docker-image.html.\n- **Newest content review: `0.8.0`** ([review](https:\u002F\u002Fgithub.com\u002Fhappy520ai\u002Funified-ai-system\u002Fblob\u002Fmaster\u002Fdocs\u002Fsecurity\u002Fmcp-image-review-0.8.0.md)),\n  read from the published layer tarballs with `node tools\u002Finspect-image-filesystem.mjs 0.8.0` - no Docker\n  engine and nothing executed. It found one thing worth knowing before you pull: the `linux\u002Farm64` tag carries\n  x86-64 native modules, including `better_sqlite3.node`, so request `--platform linux\u002Famd64`.\n- **Reviewed and pinned below: `0.4.9`.** The inspection procedure in this file\n  pins that image's recorded digests because `0.4.9` is the newest version reviewed by that procedure, which\n  exports a container filesystem with Docker. The 0.8.0 review above uses a different method and pins its own\n  digests, so the table below still describes 0.4.9 until it is rewritten against a 0.8.0 export.\n  `0.4.9` carries 9 of the fifteen names: the model-backed enhancement, knowledge\n  retrieval and workflow execution tools arrived at 0.5.0, and the three\n  governance tools at 0.8.0.\n\nDo not substitute a mutable tag for a pinned digest, and do not move the pin to a\nnewer version just because this file looks out of date: a new content review is\nrequired first, and the pinned identity is only as good as the review that backs it.\n\n## Prerequisites And Setup\n\n1. Confirm that Codex CLI and Docker are installed and Docker is running.\n2. If the nine tools are already visible, skip setup and do not register a\n   duplicate server.\n3. Explain the first stage: it downloads one reviewed platform from the\n   immutable `0.4.9` multi-platform index into Docker's cache, inspects its\n   metadata and layer history, creates but never starts a temporary container,\n   exports its root filesystem, removes that temporary container, and writes an\n   inspection inventory to a temporary directory. The reviewed platforms are\n   linux\u002Famd64 and linux\u002Farm64. Obtain explicit user approval for those download\n   and inspection changes only.\n4. After that first approval, pull the reviewed platform manifest and complete\n   the inspection. Do not execute the image or register it yet:\n\n```bash\nIMAGE='ghcr.io\u002Fhappy520ai\u002Funified-ai-system\u002Fmcp-server@sha256:751a0d32acd2d6b1da6ad9ac67987fbd1ff36ce26b7160014d8605f18b7907b3'\nPLATFORM='linux\u002Famd64' # Use linux\u002Farm64 only on a reviewed ARM64 engine.\nREVIEW_DIR=\"$(mktemp -d)\"\n\ndocker pull --platform \"$PLATFORM\" \"$IMAGE\"\ndocker image inspect \"$IMAGE\" --format 'Id={{.Id}} OS={{.Os}} Architecture={{.Architecture}} User={{json .Config.User}} Entrypoint={{json .Config.Entrypoint}} Cmd={{json .Config.Cmd}} Labels={{json .Config.Labels}}'\ndocker image history --no-trunc \"$IMAGE\" > \"$REVIEW_DIR\u002Fimage-history.txt\"\n\nREVIEW_CONTAINER=\"$(docker create --platform \"$PLATFORM\" --pull never --entrypoint \u002Fbin\u002Ftrue \"$IMAGE\")\"\ndocker export --output \"$REVIEW_DIR\u002Frootfs.tar\" \"$REVIEW_CONTAINER\"\ndocker rm \"$REVIEW_CONTAINER\"\n\ntar -tf \"$REVIEW_DIR\u002Frootfs.tar\" > \"$REVIEW_DIR\u002Frootfs-files.txt\"\nmkdir -p \"$REVIEW_DIR\u002Frootfs\"\ntar --same-permissions -xf \"$REVIEW_DIR\u002Frootfs.tar\" -C \"$REVIEW_DIR\u002Frootfs\"\nfind \"$REVIEW_DIR\u002Frootfs\u002Fapp\" -type f -print > \"$REVIEW_DIR\u002Fapp-files.txt\"\n: > \"$REVIEW_DIR\u002Fapp-links.txt\"\nwhile IFS= read -r -d '' APP_LINK; do\n  ls -ld -- \"$APP_LINK\" >> \"$REVIEW_DIR\u002Fapp-links.txt\"\ndone \u003C \u003C(find \"$REVIEW_DIR\u002Frootfs\u002Fapp\" \\( -type l -o -type f -links +1 \\) -print0)\n: > \"$REVIEW_DIR\u002Fnative-binaries.sha256\"\nwhile IFS= read -r -d '' NATIVE_BINARY; do\n  sha256sum -- \"$NATIVE_BINARY\" >> \"$REVIEW_DIR\u002Fnative-binaries.sha256\"\ndone \u003C \u003C(find \"$REVIEW_DIR\u002Frootfs\u002Fapp\" -type f -name '*.node' -print0)\nfind \"$REVIEW_DIR\u002Frootfs\" -type f \\( -perm -0100 -o -perm -0010 -o -perm -0001 \\) -print > \"$REVIEW_DIR\u002Fexecutable-files.txt\"\nfind \"$REVIEW_DIR\u002Frootfs\" -type f \\( -perm -4000 -o -perm -2000 \\) -print > \"$REVIEW_DIR\u002Fsuid-sgid-files.txt\"\nfind \"$REVIEW_DIR\u002Frootfs\u002Fapp\" -type f \\( -name '.env' -o -name '.env.*' -o -name '*.pem' -o -name '*.key' -o -name '*.p12' -o -name '*.pfx' -o -path '*\u002F.ssh\u002Fid_*' \\) -print > \"$REVIEW_DIR\u002Fcredential-like-files.txt\"\ngrep -RInHE --include='package.json' '\"(preinstall|install|postinstall|prepare|prepack|postpack)\"' \"$REVIEW_DIR\u002Frootfs\u002Fapp\" > \"$REVIEW_DIR\u002Flifecycle-hooks.txt\"\ngrep -RInE 'child_process|spawn\\(|fetch\\(|AI_GATEWAY_MCP_URL|process\\.env|writeFile|appendFile|unlink|rm\\(' \"$REVIEW_DIR\u002Frootfs\u002Fapp\u002Fpackages\u002Fmcp-server\u002Fsrc\" \"$REVIEW_DIR\u002Frootfs\u002Fapp\u002Fpackages\u002Fshared-sdk\u002Fsrc\" > \"$REVIEW_DIR\u002Fruntime-sensitive-code.txt\"\n```\n\nIf `sha256sum` is unavailable, use the platform's SHA-256 utility and preserve\nthe same report. Keep the review directory until the report is accepted; its\ndeletion is another filesystem change and requires approval for the exact path.\n\n5. Read every generated inventory and report the inspection before proceeding.\n   Compare it with the versioned\n   [image content review](https:\u002F\u002Fgithub.com\u002Fhappy520ai\u002Funified-ai-system\u002Fblob\u002Fmaster\u002Fdocs\u002Fsecurity\u002Fmcp-image-review-0.4.9.md).\n   Require OCI index digest\n   `sha256:751a0d32acd2d6b1da6ad9ac67987fbd1ff36ce26b7160014d8605f18b7907b3`.\n   For linux\u002Famd64, require manifest digest\n   `sha256:ff6cf988b01d5fb2e97aabe8e952f6a303dcffe650df5b4dcb0ba3d51ee88c06`\n   and config digest\n   `sha256:0c2c0c7b9c7fb7ca24c73d9a903bcf719b079a0b285a3a3269ee3ae059905e97`.\n   For linux\u002Farm64, require manifest digest\n   `sha256:90318b9e373820f863c1c1addc759be4b5ce186f2ecb6232ee502fad7c6613de`\n   and config digest\n   `sha256:c2047eb63fdc42bcb16d53fca17d78a4a6fb355cf6320b9aa6688e594371054f`.\n   Require source `https:\u002F\u002Fgithub.com\u002Fhappy520ai\u002Funified-ai-system`, revision\n   `342a47313927870bcc696be13c9e5fb922062dac`, version `0.4.9`, license\n   `Apache-2.0`, entrypoint `docker-entrypoint.sh`, and command\n   `node packages\u002Fmcp-server\u002Fsrc\u002Findex.js`.\n\n   Report these reviewed risks explicitly: the image uses the default root\n   user; includes Debian shell\u002Fpackage utilities and 11 base-image SUID\u002FSGID\n   files; contains 522 internal pnpm links, three native Node binaries, and eight\n   lifecycle-hook declarations; and starts a child gateway with loopback HTTP.\n   The optional `AI_GATEWAY_MCP_URL` can make an HTTP or HTTPS connection only\n   when explicitly passed. The registered command below passes no host files,\n   environment variables, or ports and disables container networking. Stop on\n   any mismatch, unexpected link, credential-like file, native binary, hook,\n   privileged file, or sensitive-code behavior.\n6. Explain the second stage: it persists a Codex MCP configuration and permits\n   Codex to launch the inspected image in a later task. Obtain a separate\n   explicit approval for registration and activation; the download approval\n   does not carry over.\n7. After that second approval, register the reviewed platform digest with\n   pulling, container networking, Linux capabilities, and privilege escalation\n   disabled, then inspect the stored configuration:\n\n```bash\nIMAGE='ghcr.io\u002Fhappy520ai\u002Funified-ai-system\u002Fmcp-server@sha256:751a0d32acd2d6b1da6ad9ac67987fbd1ff36ce26b7160014d8605f18b7907b3'\nPLATFORM='linux\u002Famd64' # Match the reviewed platform inspected above.\ncodex mcp add unified-ai-system -- docker run --rm -i --pull never --platform \"$PLATFORM\" --network none --cap-drop ALL --security-opt no-new-privileges \"$IMAGE\"\ncodex mcp get unified-ai-system --json\n```\n\n8. Restart Codex or open a new task, then use `\u002Fmcp verbose` to confirm that all\n   nine tools are available - the pinned `0.4.9` image ships nine of the fifteen\n   names the current release declares. Remove the registration when it is no longer\n   wanted:\n\n```bash\ncodex mcp remove unified-ai-system\n```\n\nRemoving the registration does not remove the pulled image from Docker's\ncache. Treat image-cache deletion as a separate host-state change and obtain\napproval before doing it.\n\n## When to Use This Skill\n\n- Use when a user asks whether Unified AI System is healthy or ready.\n- Use when a user wants a credential-free gateway chat proof.\n- Use when a user asks about the gateway's knowledge, workflow, or workforce\n  surfaces.\n- Use when a user wants evidence from the bundled MCP tools rather than a claim\n  inferred from documentation or process exit codes.\n\nDo not use this skill for generic model comparisons, unrelated MCP servers, or\ndeploying a production gateway.\n\n## Workflow\n\n1. Confirm that the `unified-ai-system` MCP tools are available in the current\n   task. If they are absent, follow the approved setup above and wait for a\n   restarted or new task.\n2. Call `gateway_health`, then `gateway_readiness`, before attempting chat.\n3. Select the narrowest additional tool that answers the request.\n4. Report returned provider, execution mode, readiness, and blockers exactly.\n5. Separate transport success from product, production-readiness, autonomy, or\n   AGI claims.\n\n## Tool Map\n\nNine of these ship in the reviewed `0.4.9` image below; the six marked 0.5.0 and\n0.8.0 are in the current release and are absent from that older image.\n\nStatus and boundaries:\n\n- `gateway_health`: gateway health, provider mode, and the real-provider safety flag\n- `gateway_readiness`: first-run readiness for chat and the local gateway runtime\n- `knowledge_readiness`: knowledge infrastructure without loading or changing data\n- `workflow_health`: the governed workflow subsystem without starting a workflow\n- `workflow_actions`: workflow action definitions without invoking any action\n- `workforce_health`: the workforce subsystem without planning or executing work\n- `workforce_agents`: configured workforce agent descriptors without dispatching them\n\nDoing work locally, still with no provider call:\n\n- `gateway_prompt_enhance`: structures a plain-language request into a prompt\n  locally, without provider credentials or provider calls\n- `gateway_prompt_enhance_llm` *(0.5.0)*: semantic rewriting through a provider\n  when one is configured, falling back to the deterministic local engine when none is\n- `knowledge_retrieve` *(0.5.0)*: keyword search over the local knowledge base,\n  returning ranked chunks with citations; calls no provider\n- `workflow_run` *(0.5.0)*: the 3-step local workflow - retrieve knowledge, compose\n  a Markdown report, write a controlled artifact; calls no provider\n- `gateway_chat`: one chat request, accepted only when the gateway proves real\n  providers are disabled\n\nThe governed Agent surface, read-only *(0.8.0)*:\n\n- `agent_governance_status`: governance status through the authenticated Gateway\n  identity; no tenant override is accepted and it fails closed when that identity is\n  not authorized for platform status\n- `agent_governance_list`: only the governed Agents visible to that tenant, with no\n  tenant or owner override argument\n- `agent_governance_describe` (takes `agentId`): one Agent as seen by that tenant;\n  cross-tenant and missing identifiers stay indistinguishable on purpose\n\nCreating, executing, revoking, approving or activating anything in that surface\nremains a human REST\u002FSDK\u002FCLI operation and is not exposed to the model at all.\n\n## Example\n\n```text\nUser: Check whether the local gateway is ready, then prove chat works safely.\n\nAgent:\n1. Call gateway_health.\n2. Call gateway_readiness.\n3. Call gateway_chat only if both results prove fake-provider mode.\n4. Report provider, model, execution mode, response, and every blocker.\n```\n\n## Safety Boundaries\n\n- Keep the credential-free local fake provider as the default.\n- Never request, read, or transmit provider credentials through this skill.\n- Do not enable or call a real provider without explicit scoped authorization.\n- Treat MCP registration, image pulls, container creation, networking, and\n  teardown as host-state changes that require informed user approval.\n- Never substitute a mutable tag, a different OCI index, or an unreviewed\n  platform manifest for the reviewed `0.4.9` identities. Keep download and\n  inspection approval separate from registration and activation approval.\n- Keep `--pull never` in the registered command. If the reviewed image is\n  absent from the local cache, fail closed and return to the first approval\n  stage.\n- Keep `--network none`, `--cap-drop ALL`, and\n  `--security-opt no-new-privileges` in the registered command.\n- Do not claim production readiness, L5 autonomy, or AGI from a healthy handshake.\n- Treat a zero exit code as transport evidence, not proof that readiness gates\n  passed.\n\n## Limitations\n\n- This skill file does not bundle the MCP server, Docker image, or Codex\n  configuration. It only operates tools supplied by the separately installed\n  official integration.\n- It does not deploy, benchmark, or certify the gateway for production use.\n- The credential-free chat tool proves only the deterministic local fake path.\n- It does not configure real providers or handle provider credentials.\n- The published MCP image requires Docker.\n- The reviewed `0.4.9` path covers linux\u002Famd64 and linux\u002Farm64. Do not activate\n  another platform image without a separate content review.\n- The image runs as the container's default root user and bundles the gateway\n  source, package-manager tooling, native dependencies, and base-image\n  SUID\u002FSGID files. The registered command drops capabilities, prevents new\n  privileges, disables networking, and leaves the image in Docker's cache.\n- Existing Codex tasks may not hot-load a newly installed MCP configuration.\n\n## Troubleshooting\n\n- If the tools are missing after approved registration, inspect\n  `codex mcp get unified-ai-system --json`, then restart Codex or start a new\n  task.\n- If readiness is blocked, report the returned blocker instead of retrying chat\n  blindly.\n- If the runtime might use a real provider, stop before chat and keep the\n  session read-only.\n\n## Additional Resources\n\n- [Unified AI System](https:\u002F\u002Fgithub.com\u002Fhappy520ai\u002Funified-ai-system)\n- [60-second Codex MCP quickstart](https:\u002F\u002Fgithub.com\u002Fhappy520ai\u002Funified-ai-system\u002Fblob\u002Fmaster\u002Fdocs\u002Fcodex-mcp-quickstart.md)\n- [MCP server guide](https:\u002F\u002Fgithub.com\u002Fhappy520ai\u002Funified-ai-system\u002Fblob\u002Fmaster\u002Fpackages\u002Fmcp-server\u002FREADME.md)\n- [MCP image content review](https:\u002F\u002Fgithub.com\u002Fhappy520ai\u002Funified-ai-system\u002Fblob\u002Fmaster\u002Fdocs\u002Fsecurity\u002Fmcp-image-review-0.4.9.md)\n",{"name":5,"license":15,"metadata":51,"description":7},{"risk":34,"tags":52,"tools":53,"author":54,"source":55,"category":56,"date_added":57,"source_repo":58,"source_type":45,"license_source":59},"ai-gateway, codex, mcp, self-hosted, governance","codex","happy520ai","https:\u002F\u002Fgithub.com\u002Fhappy520ai\u002Funified-ai-system\u002Ftree\u002Fmaster\u002Fskills\u002Funified-ai-gateway","ai-ml","2026-08-01","happy520ai\u002Funified-ai-system","https:\u002F\u002Fgithub.com\u002Fhappy520ai\u002Funified-ai-system\u002Fblob\u002Fmaster\u002FLICENSE","main","2b63f856ce19c2a797929cacabd9e0fc936812fa","plugins\u002Fhappy520ai\u002Funified-ai-system\u002Fskills\u002Funified-ai-gateway",483,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins","https:\u002F\u002Fhol.org\u002Fplugins\u002Fbest-codex-plugins","https:\u002F\u002Fgithub.com\u002Fhashgraph-online","https:\u002F\u002Fhol.org","Open-source standards and developer tools for interoperable AI agents: discovery, identity, communication, payments, privacy and security.","2026-11-13 02:16:32",{"source":8,"skillName":5,"commands":71,"prompt":81,"zipAvailable":82},[72,75,78],{"tool":73,"command":74},"npx","npx skills add hashgraph-online\u002Fawesome-codex-plugins --skill unified-ai-gateway",{"tool":76,"command":77},"bunx","bunx skills add hashgraph-online\u002Fawesome-codex-plugins --skill unified-ai-gateway",{"tool":79,"command":80},"pnpm","pnpm dlx skills add hashgraph-online\u002Fawesome-codex-plugins --skill unified-ai-gateway","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fhappy520ai\u002Funified-ai-system\u002Fskills\u002Funified-ai-gateway\u002FSKILL.md",true,[84],{"path":85,"resourceType":86,"resourceTypeDesc":87,"sizeBytes":88,"sourceHtmlUrl":89},"agents\u002Fopenai.yaml","REFERENCE","参考文档",203,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fhappy520ai\u002Funified-ai-system\u002Fskills\u002Funified-ai-gateway\u002Fagents\u002Fopenai.yaml",[91,99,106,113,140,165],{"skillNo":92,"name":93,"nameZh":6,"description":94,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":95,"categories":96,"keywords":97,"publishedAt":98},"01M746ZWACVKMZ1BJXJZHYJSSX","migrate-to-e2e","Moves an existing Playwright, Cypress, Selenium, Appium, Detox, Maestro, or TesterArmy suite to e2e one slice at a time, with the old suite running beside the new one until each ported test is proven. Keeps a MIGRATION.md progress file so any later session resumes where the last stopped. Optionally plants a defect to show both suites catch it. Use for \u002Fmigrate-to-e2e, \"migrate our tests to e2e\", \"port these Maestro flows\", \"move our TesterArmy tests\", or when a project has an e2e config and an older suite still running.","https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Ftester-army\u002Fe2e\u002Fskills\u002Fmigrate-to-e2e\u002FSKILL.md",[],[],"2026-11-11 15:11:14",{"skillNo":100,"name":101,"nameZh":6,"description":102,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":103,"categories":104,"keywords":105,"publishedAt":98},"01M746ZPTAX1Q0FBKAB1CYND7E","e2e","Writes, runs, and debugs end-to-end tests for web and mobile apps with e2e, the agentic test runner, which mixes agent.act goals with exact locator assertions on a Playwright browser or an iOS or Android device. Also drives a project's running app live over the e2e MCP server to look at a screen, check a UI change, take a screenshot or record a video, and try a locator; and runs bug bashes, parallel e2e explore runs with each bug proven by a repro test. Use when a project depends on e2e or has an e2e.config.ts, when asked for end-to-end, browser, mobile, or agentic UI tests, to check a change in the running app, to QA or bug bash an app or branch, or when an e2e run fails.","https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Ftester-army\u002Fe2e\u002Fskills\u002Fe2e\u002FSKILL.md",[],[],{"skillNo":107,"name":108,"nameZh":6,"description":109,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":110,"categories":111,"keywords":112,"publishedAt":98},"01M746ZJN784HVHMDP9MQGCSGG","create-verification-skill","Generate a project-local verify-\u003Capp> skill on top of e2e, so any coding agent can launch the app, check it is worth driving, exercise a feature the way a user does, keep evidence, and bug bash it. Writes the skill, a feature map with one explore charter per feature, and proves it once. Use for \u002Fcreate-verification-skill, \"make a verification skill for this repo\", or when a project has no scripted way to prove UI behavior.","https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Ftester-army\u002Fe2e\u002Fskills\u002Fcreate-verification-skill\u002FSKILL.md",[],[],{"skillNo":114,"name":115,"nameZh":6,"description":116,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":117,"categories":118,"keywords":119,"publishedAt":46},"01M716A6ETNB9B1SQ3GXZC75VR","deja-search","deja-vu memory — search the user's past AI coding sessions with the deja CLI. Holds what they already settled and never wrote into the repo (naming rules, banned or already-tried tools, required checks before merge, preferred values, who owns what). Use before your first edit in a task, before you call a change done or ready to merge, before you recommend a tool, library, design or reviewer, before re-debugging an error that may already be solved, when they say 'didn't we fix this before', 'what did we decide about X' or 'I already have that', and before telling them something on this machine does not exist.","https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fvshulcz\u002Fdeja-vu\u002Fskills\u002Fdeja-search\u002FSKILL.md",[],[120,121,122,123,124,125,126,127,128,129,130,131,132,133,134,135,136,137,138,139],"search","deja search","never","error","ready","memory","re","decide","fix","design","debugging","when","you","done","reviewer","io","change","cli","merge","past",{"skillNo":141,"name":142,"nameZh":6,"description":143,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":144,"categories":145,"keywords":146,"publishedAt":46},"01M716A3J2W5SENSE9VGMXB375","deja-reflect","End-of-session reflection and memory curation with deja-vu. Review what was decided, what commands broke and what repaired them, and record durable project rules or decisions with deja remember so the next session starts with them.","https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fvshulcz\u002Fdeja-vu\u002Fskills\u002Fdeja-reflect\u002FSKILL.md",[],[147,148,149,135,150,151,152,153,154,155,156,157,158,159,125,160,161,162,163,164],"deja reflect","reflect","review","with","record","github","next","https","node","Rules","decisions","brew","remember","project rules","project","session","end","reflection",{"skillNo":166,"name":167,"nameZh":6,"description":168,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":169,"categories":170,"keywords":178,"publishedAt":46},"01M715T3RR4BAYSR57HPPTJ0PF","revid","Make, edit, export and post AI videos with Revid (revid.ai) through the Revid MCP server. Use whenever the user wants a short-form or long-form video made (from an idea, a script, a URL, an article, a PDF, a tweet, a product page, their own clips, or a song), wants to find what is going viral on TikTok, YouTube, Instagram or X before making something, wants to edit or re-export an existing Revid project, generate a standalone AI image or clip, make a series episode, or publish and schedule a finished video to connected social accounts. Covers setup (hosted MCP with OAuth, or an API key), the 36 tools, credits, polling and the rules for publishing. Routes to the more specific revid-* skills when one fits.","https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Ftmakersas\u002Frevid-skills\u002Fskills\u002Frevid\u002FSKILL.md",[171,174,175],{"categoryNo":172,"categoryName":173,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000016","图像与多媒体",{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},{"categoryNo":176,"categoryName":177,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000010","后端与接口",[167,179,180,181,182,183,184,161,185,126,186,187,37,188,131,189,190,191,192,193],"URL","Video","short-form","standalone","PDF","viral","tiktok","short","social","song","image","mcp server","schedule","api","publish"]