{
  "sources": [
    {
      "url": "https://raw.githubusercontent.com/RaYYeR220/housemate/09c1a1d0a88bfa8762cee6b3ccf94271b0f172c5/core/hub.py",
      "quote": "Trust model: a satellite says ``hello``. With a valid device key it is trusted; with a\nwrong key the socket is closed with 4401; without a key it only gets a 6-digit pairing\ncode and nothing it sends is accepted until the owner approves that code out of band\n(``Hub.approve_pairing``: Desktop Rooms page or ``hermes housemate pair``). A denied device\n(``Hub.deny_pairing``) is closed with 4403 and gets no new code for ``deny_ttl_s``.\n\"\"\""
    },
    {
      "url": "https://raw.githubusercontent.com/opeblow/kata/cc59b08f1bbeeb02524c060faac04edb34de0564/README.md",
      "quote": "Kata is a plugin for [Hermes Agent](https://github.com/NousResearch/hermes-agent). It watches the tool calls your agent\nreally makes, finds the multi-step routines it keeps rediscovering across conversations, and, **only after you review\nand approve them**, compiles each one into a single deterministic tool. No model calls between steps. Every run traced.\nNothing activates without you."
    },
    {
      "url": "https://raw.githubusercontent.com/Joko-F/hermes-fine-print/7b574f1b5c8dc8eafe3ea9f87ae6b0111853f0ba/README.md",
      "quote": "- A rule that raises three times is switched off until the process restarts and shows as such in\n  `/limits self-test`; an unknown format yields nothing. Formats are pinned to v0.23.0 and the self-test\n  says when the running version differs or cannot be read."
    },
    {
      "url": "https://raw.githubusercontent.com/otsoa-dubief/tenure/f4a62855e592b9e170ab0c54d9278aabef05f370/README.md",
      "quote": "Tenure keeps Hermes' speed and adds what is missing:\n\n| Tier | What happens |\n|---|---|\n| **Trusted** (the start) | For files, memory, subagents and the usual local dev tools (npm, python, pytest, cargo, go, make, git...), the agent works as with Hermes alone: no extra question. Files it changes are saved first and can be undone for 24 hours (`/tenure undo`); anything can be flagged (`/tenure flag`). Trust is given, not earned: it reaches no tier and claims no guarantee. |"
    },
    {
      "url": "https://raw.githubusercontent.com/medomohamed22/hermes-ark/a23e529c1f636aa49d974576cdc5cb15fc8242d3/README.md",
      "quote": "**Encrypted, versioned backups of your whole Hermes — automatic, off-site, and restorable one piece at a time.**"
    },
    {
      "url": "https://raw.githubusercontent.com/robbyczgw-cla/hermes-ui-cards/c40df4d41cd5439cf4902995fd4a9e5505483f25/README.md",
      "quote": "- **Answered cards lock.** An answered card shows a receipt and locks in that window. The plugin's\n  backend records answers too, so the Webapp, the Desktop app and a second computer on the same Hermes\n  lock the card once they have checked (its buttons wait for that first check). This is eventual, not\n  exactly-once: two clients that answer before either has synced can both send; the backend keeps the\n  first answer and tells the model about the second."
    },
    {
      "url": "https://raw.githubusercontent.com/akinduroifedayo/shadow/38173bc9c696df3b1f87b5b9b60ccecd36a260ab/README.md",
      "quote": "Shadow is a Hermes plugin. It replays recent saved turns on the model you want to try and your current model. A judge compares the answers. The result leads with the numbers: turns won, lost, and tied, and the cost difference per turn. It also shows the win rate with its 95% range, speed, tool steps, and broken calls. You can inspect individual turns and switch models with undo. If the test doesn't back the switch, Shadow warns you first and switches only if you confirm."
    },
    {
      "url": "https://raw.githubusercontent.com/0xNoramiya/hypnos/0a809b9085fdeaed63fd909b927afcc41de512de/README.md",
      "quote": "During the day Hypnos watches how you work with Hermes: what you approve, what you deny, what you correct,\nand which commands pass or fail. At night it sleeps. It replays the day, dreams up variations of its\nmistakes and practises them in a sandbox, then consolidates what it learned into the weights of a small\npersonal model, a LoRA brain. In the morning the brain takes an exam. If it beats the brain it replaces,"
    },
    {
      "url": "https://raw.githubusercontent.com/breadcrumbs-agency/hermes-breadcrumb-plugin/b8975645e5174edbcc00d5dc86a78e536bce938b/README.md",
      "quote": "breadcrumbs is a memory and a context engine for the Hermes agent. It keeps every fact with its exact source.\nEvery line has an address, and one tool shows any memory at a level from 0 (handle) to 1 (all, verbatim): `crumb(id, level)`."
    },
    {
      "url": "https://raw.githubusercontent.com/miracle352/hermes-ariadne/128c95967a0761ad24d3d224441d1502c43bbb69/driver.py",
      "quote": "    surface = opts.surface or tape.header.get(\"platform\") or \"tui\"\n    if surface == \"desktop\":\n        env[\"HERMES_DESKTOP\"] = \"1\"\n    else:\n        env.pop(\"HERMES_DESKTOP\", None)\n    _install_into(scratch, opts, env)"
    },
    {
      "url": "https://raw.githubusercontent.com/merdoks/trust-ladder/fae4e9df58767e5e3832c8f65936b97868bc98fc/README.md",
      "quote": "| **1 · impossible** | A guard (regex / path glob / tool matcher) that **blocks** the call before it runs, telling the model why and what to do instead | `pre_tool_call` → `block` |\n| **2 · checked** | A check on what a write puts into matching files (or on a tool's output); a failure is appended to the tool result so the model fixes it at once | `transform_tool_result` |"
    },
    {
      "url": "https://raw.githubusercontent.com/PedValme/foreman/c15453cb2fc07fa674d6f7b548a553b3ce9e6f5d/plugin/README.md",
      "quote": "**A green check the agent did not rig.**\n\nForeman is a [Hermes Agent](https://github.com/NousResearch/hermes-agent) plugin for people who let Hermes write\ncode in a git repository. A coding agent under pressure can make the check pass by skipping a test, deleting it,\nloosening an assertion or editing what the check runs. Foreman reads the real diff before every commit, push, pull"
    },
    {
      "url": "https://raw.githubusercontent.com/Bruno-coder21/fluent/182c4301f2437e784e9022f1f38576e1d68c3f5f/README.md",
      "quote": "Hermes already checks the code it writes. **fluent checks the shell commands it runs**,\nagainst the real `--help` and `man` pages of the program installed on *your* machine,\nbefore they run."
    },
    {
      "url": "https://raw.githubusercontent.com/bettersandeep/grimoire/4824773040b131f3179e5896d71d1435908758a7/engine.py",
      "quote": "Safety model:\n- DuckDB runs in memory. CSV and Parquet files are views; SQLite tables are copied in.\n- After setup, only the attached files are readable (allowed_paths), all other file\n  access is off, extension install and load are off, and the configuration is locked,\n  so a query cannot write files, read other files, download code or undo the settings.\n- Only read statements are accepted, results are capped, and slow queries are interrupted.\n\"\"\""
    },
    {
      "url": "https://raw.githubusercontent.com/jeremiebergel/context-manager-intelligence/399468f0687c051476961e4bb045eb84263be6b1/cmi/engine.py",
      "quote": "Compiles each outgoing request (select_context): what is sent is shortened, never lost. In Hermes,\nfull-window compaction is delegated to the built-in ContextCompressor; standalone (proxy, other\nharnesses) there is no compactor. Provider/model-agnostic. No LLM call in select_context().\n\"\"\"\nfrom __future__ import annotations"
    },
    {
      "url": "https://raw.githubusercontent.com/0xbyt4/hermes-canvas/d6eeff0d6f7ec653262537bdd7bb001cc6d3c7f8/README.md",
      "quote": "Hermes Canvas is a plugin for [Hermes Agent](https://github.com/NousResearch/hermes-agent) and\nHermes Desktop. The agent draws its plan as hand-drawn boxes on a canvas next to the chat. Each\nbox is bound to files and a check command. Box status comes from hooks, not from the model: a box\nis \"doing\" when the agent writes one of its files, and \"done\" or \"failed\" when its check command\nexits. You steer the build by editing the canvas, before and during the build."
    },
    {
      "url": "https://raw.githubusercontent.com/cineglobe/hermes-gemini-live/b6c2f24354aeed1a8b5651b62944d14ed96a472d/glive_core.py",
      "quote": "* the Gemini key, read only through ``agent.secret_scope.get_secret`` (never ``os.environ``);\n* the Live model list and the ephemeral-token mint. The token LOCKS the whole session setup\n  (model, voice, system instruction, tools) server-side, so the renderer never sees the key and\n  cannot widen what the voice model may do;\n* the system instruction, personas and the tool catalog (with a ``where`` route per tool);"
    },
    {
      "url": "https://raw.githubusercontent.com/vijaygopalbalasa/hermes-necromancer/0ff7bec40aba10091a84f94025d4471984fd7386/README.md",
      "quote": "Lock the tests that define done with one command. When Hermes says it's finished, the plugin runs them itself and\nsends Hermes back with the failure named. Edits to locked tests are blocked, and a pass ends the reply with a receipt\nanyone can re-check."
    },
    {
      "url": "https://raw.githubusercontent.com/akinduroifedayo/living-spaces/e6145ccbd028edf55417822020426f4830343eb6/src/living-spaces/README.md",
      "quote": "Upkeep uses the private `cron.jobs` and `cron.executions` modules. Desktop session cleanup uses the raw `window.hermesDesktop` bridge to retain the original backend target. These integrations carry compatibility risk and catalog-review warnings. This is not an all-public-SDK or catalog-approved package."
    },
    {
      "url": "https://raw.githubusercontent.com/Pratiikpy/workroom-for-hermes/c5f6bf334c1163951e203dc7c5777faadfe49795/README.md",
      "quote": "To be able to put protected files back, the safety net reads, in memory, the bytes of every managed or protected file of a registered project (including `.env` and key files) before a shell, code or editing tool runs; those bytes are never shown to the agent and are kept in plugin data only if the agent changed the file (the agent's version, so you can recover it)."
    }
  ]
}