[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7B0DPB15FK88ECNHH3SSMY3":3,"skill-same:01M7B0DPB15FK88ECNHH3SSMY3":70},{"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":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":37,"commitSha":38,"skillDirectory":39,"forksCount":14,"repositoryHtmlUrl":40,"repositoryHomepageUrl":6,"ownerHtmlUrl":41,"ownerBlogUrl":6,"ownerDescription":42,"updatedAt":43,"install":44,"resources":57},"01M7B0DPB15FK88ECNHH3SSMY3","local-daemon-mcp-integration",null,"Wire a local Node daemon into Hermes as an MCP server.","dbillion\u002Fhermes-skills-public","dbillion","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F49393855?v=4","01M7AVH8QPGCM8WKWS22TRT0T6","Oludayo Adeoye","01M7AVH8RD1JJWA0YS9CFEWAV9",0,"MIT","https:\u002F\u002Fgithub.com\u002Fdbillion\u002Fhermes-skills-public\u002Fblob\u002Fmaster\u002Fskills\u002Fmcp\u002Flocal-daemon-mcp-integration\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: local-daemon-mcp-integration\ndescription: \"Wire a local Node daemon into Hermes as an MCP server.\"\nversion: 1.0.0\nauthor: Hermes Agent\nlicense: MIT\nplatforms: [linux, macos]\nmetadata:\n  hermes:\n    tags: [MCP, local-daemon, node, integration, troubleshooting]\n---\n\n# Local Daemon → Hermes MCP Integration\n\nClass of task: a developer tool ships as a **local daemon** + a **stdio MCP\nserver that proxies to it**. You register the MCP server in\n`~\u002F.hermes\u002Fconfig.yaml` so Hermes gets its tools. This skill covers the full\nflow and the two failure modes that reliably bite: native-module\u002FNode-ABI\nmismatch and `write_file`\u002F`patch` following symlinks into the real entrypoint.\n\nRead this alongside the `native-mcp` skill (Hermes' built-in MCP client).\n\n## When to use\n- \"add \u003Ctool>'s MCP server to Hermes\", \"mcp install hermes\", \"wire \u003Cdaemon> into\n  my agent\", or any local-first tool whose MCP entrypoint spawns a stdio server\n  that talks to a running daemon.\n- The daemon is a Node monorepo \u002F pnpm workspace build (Open Design, etc.).\n\n## Flow\n\n### 1. Confirm the real CLI (binary-name collision)\nMany tools expose a CLI name that collides with a system binary. Open Design's\nCLI is `od`, but `\u002Fusr\u002Fbin\u002Fod` is GNU coreutils' octal dump. A config\n`command: \"od\"` launches the wrong binary and hangs silently.\n```\ncommand -v od && od --version      # expect GNU coreutils, NOT the tool\ncommand -v \u003Calias> && \u003Calias> --help\n```\nUse the non-colliding alias (e.g. `opendesign`) and, better, the resolved\nnode-binary + entry script from step 3.\n\n### 2. The MCP server is a proxy — start the daemon first\n`od mcp` (and similar) spawns a stdio MCP server that forwards tool calls to the\nrunning daemon (e.g. `http:\u002F\u002F127.0.0.1:7456`). Tools return nothing \u002F hang\nunless the daemon is up. Start it, then verify:\n```\ncurl -s -o \u002Fdev\u002Fnull -w \"%{http_code}\\n\" --max-time 5 http:\u002F\u002F127.0.0.1:7456\u002Fapi\u002Fhealth\n# expect 200\n```\n\n### 3. Get the byte-exact launch spec — never hand-guess\nTooling usually resolves the correct command from the running daemon. Use it:\n```\n\u003Ccli> mcp install hermes --dry-run --json     # or GET \u002Fapi\u002Fmcp\u002Finstall-info\n```\nThe returned `launchSpec` is authoritative, e.g.:\n```\ncommand: \u003Cnode24>\u002Fbin\u002Fnode\nargs:    [\u003Crepo>\u002Fapps\u002Fdaemon\u002Fdist\u002Fcli.js, mcp, --daemon-url, http:\u002F\u002F127.0.0.1:7456]\nenv:     OD_DATA_DIR: \u003Crepo>\u002F.od\n```\nNote: the tool may decline to edit Hermes config itself and print \"manual setup\nrequired — add this under your Hermes MCP server configuration by hand.\" That's\nexpected; splice it yourself in step 4.\n\n### 4. Insert into config.yaml (surgical — never yaml.safe_dump)\n`~\u002F.hermes\u002Fconfig.yaml` cannot be edited with `patch`\u002F`write_file` (agent\nwrite-block), and `yaml.safe_dump` strips comments and reorders keys. Insert via\na Python script that splices raw YAML text before the next top-level key\n(e.g. `plugins:`), preserving comments\u002Forder. Back up first:\n```\ncp ~\u002F.hermes\u002Fconfig.yaml ~\u002F.hermes\u002Fconfig.yaml.bak.$(date +%s)\n```\nBlock shape:\n```\n  open-design:\n    command: \"\u003Cnode24>\u002Fbin\u002Fnode\"\n    args:\n      - \"\u003Crepo>\u002Fapps\u002Fdaemon\u002Fdist\u002Fcli.js\"\n      - \"mcp\"\n      - \"--daemon-url\"\n      - \"http:\u002F\u002F127.0.0.1:7456\"\n    env:\n      OD_DATA_DIR: \"\u003Crepo>\u002F.od\"\n```\nValidate: `python3 -c \"import yaml; yaml.safe_load(open('\u002Fhome\u002Fdeeone\u002F.hermes\u002Fconfig.yaml'))\"`.\nAlso confirm the daemon is still the sole listener on the port (ss -tlnp | grep\n7456) — avoid duplicate daemons.\n\n### 5. Verify with a real stdio JSON-RPC handshake\nConfig edit ≠ proof it runs. Spawn the exact command with the spec's env, send\n`initialize` then `tools\u002Flist` over **newline-delimited JSON** (one object per\nline — NOT LSP `Content-Length` framing, which makes the server hang). A live\n`tools\u002Fcall` returning real data is the gold standard. See\n`scripts\u002Fverify-mcp-stdio.py`.\n\n### 6. Make tools callable\nMCP tools load at agent startup. Restart Hermes (relaunch, or `\u002Freset` in a\nchat) to expose `mcp_\u003Cserver>_*` tools inline. A reload notification may fire\nwithout refreshing this session's callable tool list — that's a presentation\nquirk, not a config problem; a fresh session exposes them.\n\n## Pitfalls (transferable)\n\n### Native module \u002F Node ABI mismatch → ERR_DLOPEN_FAILED\nA daemon built on Node 24 with a native module (better-sqlite3, etc.) crashes\nunder a different Node (e.g. v25) with\n`NODE_MODULE_VERSION 137 ... requires NODE_MODULE_VERSION 141`.\nFix:\n1. Rebuild the native module against the Node you will pin:\n   `PATH=\u003Cnode24>\u002Fbin:$PATH pnpm rebuild better-sqlite3` (from repo root);\n   confirm with `node -e \"require('\u003Cmodule>')\"` under that Node.\n2. Pin the launcher to that Node: replace the alias\u002Fsymlink CLI with a wrapper\n   that prepends the Node bin dir to PATH before `exec node \u003Centry>`.\n\n### write_file \u002F patch follow symlinks — trash the symlink first\n`write_file` and `patch` RESOLVE symlinks and overwrite the target file. If the\nlauncher is a symlink chain (`~\u002F.local\u002Fbin\u002Fopendesign` → `node_modules\u002F.bin\u002Fod`\n→ `apps\u002Fdaemon\u002Fbin\u002Fod.mjs`), writing to the launcher path corrupts the REAL\nentrypoint. Rule: move the symlink to trash first, then write the real wrapper\nfile. (This has recurred across sessions — standing rule for launcher\u002Fentrypoint\nedits: trash symlink, then write file.)\n\n## Worked example: Open Design\nSee `references\u002Fopen-design-example.md` for the end-to-end Open Design flow\n(164 skills, 150 plugins, the Hermes agent adapter via ACP, and the two\ndirections of integration).\n\n## Support files\n- `scripts\u002Fverify-mcp-stdio.py` — runnable JSON-RPC handshake verifier\n  (`initialize` + `tools\u002Flist` over newline-delimited JSON). Use it to prove any\n  stdio MCP server actually works before declaring success.\n- `references\u002Fopen-design-example.md` — concrete Open Design paths, ports, the\n  byte-exact launch spec, and the Hermes ACP agent adapter.\n","# Local Daemon → Hermes MCP Integration\n\nClass of task: a developer tool ships as a **local daemon** + a **stdio MCP\nserver that proxies to it**. You register the MCP server in\n`~\u002F.hermes\u002Fconfig.yaml` so Hermes gets its tools. This skill covers the full\nflow and the two failure modes that reliably bite: native-module\u002FNode-ABI\nmismatch and `write_file`\u002F`patch` following symlinks into the real entrypoint.\n\nRead this alongside the `native-mcp` skill (Hermes' built-in MCP client).\n\n## When to use\n- \"add \u003Ctool>'s MCP server to Hermes\", \"mcp install hermes\", \"wire \u003Cdaemon> into\n  my agent\", or any local-first tool whose MCP entrypoint spawns a stdio server\n  that talks to a running daemon.\n- The daemon is a Node monorepo \u002F pnpm workspace build (Open Design, etc.).\n\n## Flow\n\n### 1. Confirm the real CLI (binary-name collision)\nMany tools expose a CLI name that collides with a system binary. Open Design's\nCLI is `od`, but `\u002Fusr\u002Fbin\u002Fod` is GNU coreutils' octal dump. A config\n`command: \"od\"` launches the wrong binary and hangs silently.\n```\ncommand -v od && od --version      # expect GNU coreutils, NOT the tool\ncommand -v \u003Calias> && \u003Calias> --help\n```\nUse the non-colliding alias (e.g. `opendesign`) and, better, the resolved\nnode-binary + entry script from step 3.\n\n### 2. The MCP server is a proxy — start the daemon first\n`od mcp` (and similar) spawns a stdio MCP server that forwards tool calls to the\nrunning daemon (e.g. `http:\u002F\u002F127.0.0.1:7456`). Tools return nothing \u002F hang\nunless the daemon is up. Start it, then verify:\n```\ncurl -s -o \u002Fdev\u002Fnull -w \"%{http_code}\\n\" --max-time 5 http:\u002F\u002F127.0.0.1:7456\u002Fapi\u002Fhealth\n# expect 200\n```\n\n### 3. Get the byte-exact launch spec — never hand-guess\nTooling usually resolves the correct command from the running daemon. Use it:\n```\n\u003Ccli> mcp install hermes --dry-run --json     # or GET \u002Fapi\u002Fmcp\u002Finstall-info\n```\nThe returned `launchSpec` is authoritative, e.g.:\n```\ncommand: \u003Cnode24>\u002Fbin\u002Fnode\nargs:    [\u003Crepo>\u002Fapps\u002Fdaemon\u002Fdist\u002Fcli.js, mcp, --daemon-url, http:\u002F\u002F127.0.0.1:7456]\nenv:     OD_DATA_DIR: \u003Crepo>\u002F.od\n```\nNote: the tool may decline to edit Hermes config itself and print \"manual setup\nrequired — add this under your Hermes MCP server configuration by hand.\" That's\nexpected; splice it yourself in step 4.\n\n### 4. Insert into config.yaml (surgical — never yaml.safe_dump)\n`~\u002F.hermes\u002Fconfig.yaml` cannot be edited with `patch`\u002F`write_file` (agent\nwrite-block), and `yaml.safe_dump` strips comments and reorders keys. Insert via\na Python script that splices raw YAML text before the next top-level key\n(e.g. `plugins:`), preserving comments\u002Forder. Back up first:\n```\ncp ~\u002F.hermes\u002Fconfig.yaml ~\u002F.hermes\u002Fconfig.yaml.bak.$(date +%s)\n```\nBlock shape:\n```\n  open-design:\n    command: \"\u003Cnode24>\u002Fbin\u002Fnode\"\n    args:\n      - \"\u003Crepo>\u002Fapps\u002Fdaemon\u002Fdist\u002Fcli.js\"\n      - \"mcp\"\n      - \"--daemon-url\"\n      - \"http:\u002F\u002F127.0.0.1:7456\"\n    env:\n      OD_DATA_DIR: \"\u003Crepo>\u002F.od\"\n```\nValidate: `python3 -c \"import yaml; yaml.safe_load(open('\u002Fhome\u002Fdeeone\u002F.hermes\u002Fconfig.yaml'))\"`.\nAlso confirm the daemon is still the sole listener on the port (ss -tlnp | grep\n7456) — avoid duplicate daemons.\n\n### 5. Verify with a real stdio JSON-RPC handshake\nConfig edit ≠ proof it runs. Spawn the exact command with the spec's env, send\n`initialize` then `tools\u002Flist` over **newline-delimited JSON** (one object per\nline — NOT LSP `Content-Length` framing, which makes the server hang). A live\n`tools\u002Fcall` returning real data is the gold standard. See\n`scripts\u002Fverify-mcp-stdio.py`.\n\n### 6. Make tools callable\nMCP tools load at agent startup. Restart Hermes (relaunch, or `\u002Freset` in a\nchat) to expose `mcp_\u003Cserver>_*` tools inline. A reload notification may fire\nwithout refreshing this session's callable tool list — that's a presentation\nquirk, not a config problem; a fresh session exposes them.\n\n## Pitfalls (transferable)\n\n### Native module \u002F Node ABI mismatch → ERR_DLOPEN_FAILED\nA daemon built on Node 24 with a native module (better-sqlite3, etc.) crashes\nunder a different Node (e.g. v25) with\n`NODE_MODULE_VERSION 137 ... requires NODE_MODULE_VERSION 141`.\nFix:\n1. Rebuild the native module against the Node you will pin:\n   `PATH=\u003Cnode24>\u002Fbin:$PATH pnpm rebuild better-sqlite3` (from repo root);\n   confirm with `node -e \"require('\u003Cmodule>')\"` under that Node.\n2. Pin the launcher to that Node: replace the alias\u002Fsymlink CLI with a wrapper\n   that prepends the Node bin dir to PATH before `exec node \u003Centry>`.\n\n### write_file \u002F patch follow symlinks — trash the symlink first\n`write_file` and `patch` RESOLVE symlinks and overwrite the target file. If the\nlauncher is a symlink chain (`~\u002F.local\u002Fbin\u002Fopendesign` → `node_modules\u002F.bin\u002Fod`\n→ `apps\u002Fdaemon\u002Fbin\u002Fod.mjs`), writing to the launcher path corrupts the REAL\nentrypoint. Rule: move the symlink to trash first, then write the real wrapper\nfile. (This has recurred across sessions — standing rule for launcher\u002Fentrypoint\nedits: trash symlink, then write file.)\n\n## Worked example: Open Design\nSee `references\u002Fopen-design-example.md` for the end-to-end Open Design flow\n(164 skills, 150 plugins, the Hermes agent adapter via ACP, and the two\ndirections of integration).\n\n## Support files\n- `scripts\u002Fverify-mcp-stdio.py` — runnable JSON-RPC handshake verifier\n  (`initialize` + `tools\u002Flist` over newline-delimited JSON). Use it to prove any\n  stdio MCP server actually works before declaring success.\n- `references\u002Fopen-design-example.md` — concrete Open Design paths, ports, the\n  byte-exact launch spec, and the Hermes ACP agent adapter.\n",{"name":5,"author":24,"license":15,"version":25,"metadata":26,"platforms":34,"description":7},"Hermes Agent","1.0.0",{"hermes":27},{"tags":28},[29,30,31,32,33],"MCP","local-daemon","node","integration","troubleshooting",[35,36],"linux","macos","master","f341c46fed021656175d5c9ccc67a72126e50bfb","skills\u002Fmcp\u002Flocal-daemon-mcp-integration","https:\u002F\u002Fgithub.com\u002Fdbillion\u002Fhermes-skills-public","https:\u002F\u002Fgithub.com\u002Fdbillion","I enjoy writing codes,teaching programming, solving problems, tasted automation in Python and never looked back😋😋😋💞","2026-10-14 03:31:55",{"source":8,"skillName":5,"commands":45,"prompt":55,"zipAvailable":56},[46,49,52],{"tool":47,"command":48},"npx","npx skills add dbillion\u002Fhermes-skills-public --skill local-daemon-mcp-integration",{"tool":50,"command":51},"bunx","bunx skills add dbillion\u002Fhermes-skills-public --skill local-daemon-mcp-integration",{"tool":53,"command":54},"pnpm","pnpm dlx skills add dbillion\u002Fhermes-skills-public --skill local-daemon-mcp-integration","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fdbillion\u002Fhermes-skills-public\u002Fblob\u002Fmaster\u002Fskills\u002Fmcp\u002Flocal-daemon-mcp-integration\u002FSKILL.md",true,[58,64],{"path":59,"resourceType":60,"resourceTypeDesc":61,"sizeBytes":62,"sourceHtmlUrl":63},"references\u002Fopen-design-example.md","REFERENCE","参考文档",2986,"https:\u002F\u002Fgithub.com\u002Fdbillion\u002Fhermes-skills-public\u002Fblob\u002Fmaster\u002Fskills\u002Fmcp\u002Flocal-daemon-mcp-integration\u002Freferences\u002Fopen-design-example.md",{"path":65,"resourceType":66,"resourceTypeDesc":67,"sizeBytes":68,"sourceHtmlUrl":69},"scripts\u002Fverify-mcp-stdio.py","SCRIPT","脚本",2866,"https:\u002F\u002Fgithub.com\u002Fdbillion\u002Fhermes-skills-public\u002Fblob\u002Fmaster\u002Fskills\u002Fmcp\u002Flocal-daemon-mcp-integration\u002Fscripts\u002Fverify-mcp-stdio.py",[71,78,85,94,101,108],{"skillNo":72,"name":73,"nameZh":6,"description":74,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":75,"categories":76,"keywords":77,"publishedAt":19},"01M7B25T0V7RBKES86Y3RTXQRN","blocked-page-recovery","Use when a fetch fails: 403\u002F429, paywall, WAF, bot wall.","https:\u002F\u002Fgithub.com\u002Fdbillion\u002Fhermes-skills-public\u002Fblob\u002Fmaster\u002Fskills\u002Fweb\u002Fblocked-page-recovery\u002FSKILL.md",[],[],{"skillNo":79,"name":80,"nameZh":6,"description":81,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":82,"categories":83,"keywords":84,"publishedAt":19},"01M7B25DSBBMZRHVD28MAX7N5X","gemini-vision-fallback","Vision fallback: Gemini REST API if vision_analyze blocked.","https:\u002F\u002Fgithub.com\u002Fdbillion\u002Fhermes-skills-public\u002Fblob\u002Fmaster\u002Fskills\u002Fvision\u002Fgemini-vision-fallback\u002FSKILL.md",[],[],{"skillNo":86,"name":87,"nameZh":6,"description":88,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":89,"licenseFileUrl":90,"sourceHtmlUrl":91,"categories":92,"keywords":93,"publishedAt":19},"01M7B22XW9QXT2ZWWXVZDFWYZP","twitter-cli","Use twitter-cli for ALL Twitter\u002FX operations — reading tweets, posting, replying, quoting, liking, retweeting, following, searching, user lookups. Invoke whenever user requests any Twitter interaction.","Apache-2.0","https:\u002F\u002Fgithub.com\u002Fdbillion\u002Fhermes-skills-public\u002Fblob\u002Fmaster\u002Fskills\u002Ftwitter-cli\u002FLICENSE","https:\u002F\u002Fgithub.com\u002Fdbillion\u002Fhermes-skills-public\u002Fblob\u002Fmaster\u002Fskills\u002Ftwitter-cli\u002FSKILL.md",[],[],{"skillNo":95,"name":96,"nameZh":6,"description":97,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":98,"categories":99,"keywords":100,"publishedAt":19},"01M7B22QMWY1B4F7BC114RG4S6","prism-scan","Structural analysis through dynamically generated cognitive lenses. Generates the optimal analytical lens for the specific code\u002Fartifact, then executes it. Finds conservation laws, structural invariants, and concrete bugs that vanilla analysis misses. Use on any code file, system design, or text artifact.","https:\u002F\u002Fgithub.com\u002Fdbillion\u002Fhermes-skills-public\u002Fblob\u002Fmaster\u002Fskills\u002Fts-superhermes-prism-scan\u002FSKILL.md",[],[],{"skillNo":102,"name":103,"nameZh":6,"description":104,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":105,"categories":106,"keywords":107,"publishedAt":19},"01M7B22JA5XCAP1ACT6V50WYR1","prism-reflect","Constraint transparency: analyzes an artifact structurally, then analyzes what its own analysis concealed. Produces a conservation law AND a constraint report showing what was maximized, what was sacrificed, and what to investigate next. The only AI skill that knows what it can't see.","https:\u002F\u002Fgithub.com\u002Fdbillion\u002Fhermes-skills-public\u002Fblob\u002Fmaster\u002Fskills\u002Fts-superhermes-prism-reflect\u002FSKILL.md",[],[],{"skillNo":109,"name":110,"nameZh":6,"description":111,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":112,"categories":113,"keywords":114,"publishedAt":19},"01M7B22EB15M75M4X3C52YGDCC","prism-full","Full Prism: multi-pass structural analysis with mandatory adversarial self-correction. Designs custom analytical passes, executes them with chaining, then attacks its own findings before synthesizing. Use for maximum depth on important code or artifacts.","https:\u002F\u002Fgithub.com\u002Fdbillion\u002Fhermes-skills-public\u002Fblob\u002Fmaster\u002Fskills\u002Fts-superhermes-prism-full\u002FSKILL.md",[],[]]