feat: preview Hermes and OpenClaw imports

CHANGELOG.md

@@ -7,6 +7,15 @@ The format follows [Keep a Changelog](https://keepachangelog.com/en/1.1.0/).
 The 0.20.0 beta line follows the frozen operator contract; earlier alpha releases
 remain historical snapshots. Beta does not mean error-free operation.
 
+## [Unreleased]
+
+### Added
+
+- Add an explicit, read-only `talos import ... --dry-run` preview for portable
+  Hermes and OpenClaw profile text, memory and Agent Skills. The preview writes
+  nothing, never activates imported instructions, and excludes secrets,
+  permissions, executable configuration, schedules, sessions and databases.
+
 ## [0.20.0-rc.2] — 2026-11-01
 
 ### Security

CLAUDE.md

@@ -14,7 +14,7 @@ preferences live in `USER.md`. All three are operator-owned prompt state and rel
 |---|---|
 | Gate path | `policy.py`, **996 lines** — has to stay readable in one sitting |
 | Tools | **37**, every one gated |
-| Suites | **3456** collected tests · **277** adversarial · 44 end-to-end |
+| Suites | **3477** collected tests · **277** adversarial · 44 end-to-end |
 | Home | <https://talos-agent.ch> · docs at `/docs/` |
 | Repository | `talos-kernel/talos` is the public source tree |
 
@@ -52,6 +52,10 @@ automatically; do not overwrite operator state with repository templates. Homebr
 installations update through `brew upgrade talos-kernel/tap/talos`; `talos update`
 refuses them. Source installations retain their own profile and update path.
 
+`talos import hermes|openclaw --source <path> --dry-run` can explicitly preview a
+small portable allowlist from another agent home. It changes neither home, imports
+no authority or secrets, and does not activate the persistent text or Agent Skills.
+
 The source branch can contain fixes newer than the latest signed release. Check
 `CHANGELOG.md`, the release tag and `site/status.json` before claiming a fix is shipped;
 a source commit does not change an existing signed archive.
@@ -313,7 +317,7 @@ repairs; corrections never replay an effect or grant permission. Keep those prop
 ```bash
 python3 -m venv .venv && . .venv/bin/activate && pip install --require-hashes -r requirements.lock -r requirements-dev.lock
 
-python -m pytest tests/ -q   # 3456 collected tests; platform-specific skips are explicit
+python -m pytest tests/ -q   # 3477 collected tests; platform-specific skips are explicit
 python redteam.py            # 277 adversarial cases — mandatory for any kernel change
 python e2e.py                # 44 cases against a real model (costs tokens and time)
 python -m talos --once       # single cycle, for diagnosis

README.md

@@ -21,10 +21,10 @@
 </p>
 
 <p align="center">
-  <!-- ⚠️ Bewusst „tests“, nicht „passing“: die Zahl kommt aus dem Einsammeln (3456).
+  <!-- ⚠️ Bewusst „tests“, nicht „passing“: die Zahl kommt aus dem Einsammeln (3477).
        Plattformabhaengige Sandbox- und Repository-Pruefungen koennen uebersprungen werden;
        `test_site_claims` prueft deshalb die gesammelte Zahl statt ein Umgebungsresultat. -->
-  <img src="https://img.shields.io/badge/tests-3456-2e7d32.svg" alt="Tests">
+  <img src="https://img.shields.io/badge/tests-3477-2e7d32.svg" alt="Tests">
   <img src="https://img.shields.io/badge/red%20team-277%2F277-2e7d32.svg" alt="Red team">
   <img src="https://img.shields.io/badge/gate%20path-996%20lines-8a4318.svg" alt="Gate path">
   <img src="https://img.shields.io/badge/tools-37%20gated-8a4318.svg" alt="Tools">
@@ -258,6 +258,13 @@ section atomically with mode 0600. It prints the configuration path and the next
 Use `talos doctor` for missing prerequisites and `talos help <command>` for command help.
 A running service needs a restart after changing its file settings.
 
+Coming from Hermes or OpenClaw? `talos import` produces a read-only, redacted plan for
+portable profile files, memory documents and Agent Skills. It does not copy anything in
+this release. Secrets, approvals, command allowlists, executable plugins and MCP commands,
+schedules, sessions and databases are skipped without reading their values. Every proposed
+skill remains inactive until a later, separately reviewed activation path passes Talos'
+existing approval boundary. See [the migration preview contract](docs/import-agents.md).
+
 Later, a single part can be redone without repeating the rest:
 
 ```bash
@@ -271,6 +278,9 @@ python -m talos config set TALOS_MODEL_OVERRIDES '{"local-model": {"context_wind
                                          # a window or a price the catalogue does not know —
                                          # never a provider, url or key; shown as your word
 python -m talos models --refresh         # force the same provider-catalogue sync now
+python -m talos import list              # supported agent-home previews
+python -m talos import hermes --source "$HOME/.hermes" --dry-run
+python -m talos import openclaw --source "$HOME/.openclaw" --dry-run
 python -m talos status                   # what it did last
 python -m talos chat                     # a session here; approvals at a terminal
 python -m talos ask "how many …?"        # one turn from here, answer on stdout
@@ -882,13 +892,14 @@ command centre.
 `/allowed` `/revoke` · `/log` `/undo` `/policy` `/autonomy` `/tools` `/whoami` `/version` ·
 `/usage` `/model` `/eval` `/reasoning` `/debug`
 
-**On the command line** — thirteen, each answering a question an operator actually asks:
+**On the command line** — each answering a question an operator actually asks:
 
 | | |
 |---|---|
 | `chat` | a session here; approvals at a terminal |
 | `ask "…"` | one turn, for scripts and cron |
 | `setup` · `doctor` · `config` · `models` | set it up, and find out why something won't work |
+| `import list` · `import <hermes\|openclaw> --source … --dry-run` | preview portable agent data without writing or activating it |
 | `status` · `events` · `why <id>` | what it did, and why that was allowed or refused |
 | `report` · `review` | a record someone else can read · what this installation should change |
 | `update` · `version` | signature checked, tests first |
@@ -925,6 +936,7 @@ sitting — a gate you cannot read is not a gate.
 | `vision.py` / `hearing.py` / `speech.py` / `frames.py` | reading a picture, hearing a recording, speaking, one still out of a video — ordinary READ/WRITE with a target |
 | `cli.py` / `doctor.py` / `configcli.py` / `schema.py` | the subcommands: diagnose, read and change settings — the schema decides what may be written, and the allowlist and the network exceptions never are |
 | `askcli.py` | `talos ask` — one turn from a script, as a channel with no special right |
+| `import_agents.py` | bounded, redacted, zero-write migration previews; imported text remains inactive |
 | `models.py` / `model_catalog_sync.py` | provider-bound live model lists; complete snapshots may replace membership, incomplete ones only add, and failures keep the last good cache |
 | `updater.py` | update beside the old tree, both suites in the new one, switch only if green |
 | `eventlog.py` / `memory.py` / `usage.py` | durable log, conversation memory, metering |

docs/import-agents.md

@@ -0,0 +1,40 @@
+# Import preview from Hermes or OpenClaw
+
+`talos import` answers one question: which portable parts of an existing agent home could
+be reviewed for Talos? The current command is deliberately a dry run. It creates no target
+directory, backup, report or temporary file and never changes the source.
+
+```bash
+talos import list
+talos import hermes --source "$HOME/.hermes" --dry-run
+talos import openclaw --source "$HOME/.openclaw" --dry-run
+```
+
+The source path is explicit. Talos accepts an absolute path after shell expansion and
+refuses a source or target boundary containing a symlink. Input files must be bounded,
+regular, singly linked UTF-8 files. The preview contains relative paths, byte counts and
+SHA-256 digests, never file bodies.
+
+## What the preview considers
+
+- `SOUL.md`, `AGENTS.md`, `USER.md` and `MEMORY.md`, where that source stores them.
+- Agent Skills with a valid `SKILL.md`, including one category-directory layer.
+- Existing staging destinations as conflicts. The preview never overwrites them.
+
+Every candidate points below `data/import-staging/<source>/`. `staged` means “eligible for
+operator review”, not active or trusted. Imported text can contain hostile instructions.
+It grants no permission, and a skill must still pass Talos' managed-skill approval path
+before it can become active.
+
+## What is always skipped
+
+- API keys, tokens, OAuth state, `.env`, `auth.json` and auth profiles.
+- Command allowlists, approvals, standing permissions and autonomy settings.
+- Plugins, hooks, MCP commands and other executable configuration.
+- Schedules, cron jobs, plans, sessions, logs and databases.
+
+Credentials are configured again through Talos' own setup flow. Permissions are never
+migrated. External connector grants must be paired separately for Talos and are never
+copied from Hermes or OpenClaw. A future apply command needs its own backup, atomic
+publication, exact-item confirmation, read-back and rollback contract; this preview does
+not imply that contract already exists.

site/console.html

@@ -378,7 +378,7 @@ <h1 id="main-content" tabindex="-1">Witness one deed from door to ledger.</h1>
       <a class="btn ghost" href="#measured">This candidate, measured</a>
     </div>
     <div class="meter-row">
-      <span><b>3456</b>automated tests</span>
+      <span><b>3477</b>automated tests</span>
       <span><b>277/277</b>adversarial</span>
       <span><b>996</b>lines · gate</span>
       <span><b>37</b>gated tools</span>
@@ -554,7 +554,7 @@ <h2 class="rv">Numbers this candidate must ship.</h2>
 
     <div class="stats" id="releaseStats">
       <div class="stat rv"><div class="n" data-f="version">0.20.0-rc.2</div><div class="l">release candidate</div><div class="sub-l" data-f="installer_version">public installer v0.20.0-rc.2</div></div>
-      <div class="stat rv"><div class="n" data-f="tests">3456</div><div class="l">automated tests, collected by pytest</div></div>
+      <div class="stat rv"><div class="n" data-f="tests">3477</div><div class="l">automated tests, collected by pytest</div></div>
       <div class="stat rv"><div class="n" data-f="redteam_faelle">277</div><div class="l">adversarial cases, all behaving</div></div>
       <div class="stat rv"><div class="n" data-f="kernel_zeilen">996</div><div class="l">lines in policy.py, readable in one sitting</div></div>
       <div class="stat rv"><div class="n" data-f="werkzeuge">37</div><div class="l">tools, every one through the gate</div></div>

site/docs/index.html

@@ -244,7 +244,7 @@ <h1 id="main-content" tabindex="-1">Talos AI agent: the field manual</h1>
     <p class="lede">That one decision shapes the rest of this manual. It is why the
     command list is short, and why a few obvious conveniences are missing on purpose.</p>
     <div class="spec-row">
-      <div class="spec"><b>3456</b>automated tests</div>
+      <div class="spec"><b>3477</b>automated tests</div>
       <div class="spec"><b>277/277</b>adversarial</div>
       <div class="spec"><b>996</b>gate lines</div>
       <div class="spec"><b>37</b>gated tools</div>

site/index.html

@@ -639,7 +639,7 @@ <h2 class="rv">Measured, with the path stated</h2>
     <p class="sub rv">None of this asks for belief. The installer runs the unit and
     adversarial suites in front of you; the real-model cases are a separate opt-in E2E run.</p>
     <p class="source-caption">Release-candidate source and signed release. Counts describe the source test suite, not speed or real-world success rates; signed release contents follow their tag.</p><div class="stats">
-      <div class="stat rv"><div class="n cnt" data-n="3456">0</div><div class="l">automated tests collected from the source tree</div></div>
+      <div class="stat rv"><div class="n cnt" data-n="3477">0</div><div class="l">automated tests collected from the source tree</div></div>
       <div class="stat rv"><div class="n cnt" data-n="277">0</div><div class="l">adversarial cases trying to pass the kernel</div></div>
       <div class="stat rv"><div class="n cnt" data-n="44">0</div><div class="l">opt-in end-to-end cases against a real model</div></div>
       <div class="stat rv"><div class="n cnt" data-n="996">0</div><div class="l">lines in the file that decides every verdict</div></div>

site/llms.txt

@@ -6,7 +6,7 @@
 
 - Permission kernel: 996 lines of Python, the only code that decides
 - 37 tools, all gated; unattended runs cannot escalate
-- 3456 tests collected in the source tree; 277/277 red-team cases passed in source verification. Release contents follow their own signed tag.
+- 3477 tests collected in the source tree; 277/277 red-team cases passed in source verification. Release contents follow their own signed tag.
 - MCP servers run only inside a UID-separated worker, from an operator-owned registry — no marketplace, no third-party code in the agent process
 - Channels: Telegram, mail (IMAP pull), CLI, WhatsApp via an operator-controlled broker — every channel fetches; no inbound port, no webhook
 - Updates: signed tarballs, verified before anything is unpacked; dependencies pinned and hashed, installed with --require-hashes

site/redteam/index.html

@@ -224,7 +224,7 @@ <h1 id="main-content" tabindex="-1">277 ways to break an agent.</h1>
       <div class="spec"><b class="red">25</b>attack surfaces</div>
       <div class="spec"><b>996</b>kernel lines</div>
       <div class="spec"><b>37</b>gated tools</div>
-      <div class="spec"><b>3456</b>automated tests</div>
+      <div class="spec"><b>3477</b>automated tests</div>
       <div class="spec"><b class="red">0</b>passes sold as safety</div>
     </div>
   </div>

site/registry/index.html

@@ -239,7 +239,7 @@ <h1 id="main-content" tabindex="-1">The MCP registry</h1>
     guide with the full protocol is <a href="https://github.com/talos-kernel/talos/blob/master/docs/claude-worker.md">docs/claude-worker.md</a>.
     A curated starting point ships as <code>examples/mcp-servers.json</code>.</p>
     <div class="spec-row">
-      <div class="spec"><b>3456</b>automated tests</div>
+      <div class="spec"><b>3477</b>automated tests</div>
       <div class="spec"><b>277/277</b>adversarial</div>
       <div class="spec"><b>996</b>gate lines</div>
       <div class="spec"><b>37</b>gated tools</div>

site/setup/index.html

@@ -163,7 +163,7 @@ <h1 id="main-content" tabindex="-1" class="rv">This address moved.</h1>
       <a class="btn ghost" href="/">Return to the keep</a>
     </div>
     <div class="spec-row rv">
-      <div class="spec"><b>3456</b>automated tests</div>
+      <div class="spec"><b>3477</b>automated tests</div>
       <div class="spec"><b>277/277</b>adversarial</div>
       <div class="spec"><b>996</b>gate lines</div>
       <div class="spec"><b>37</b>gated tools</div>

site/status.json

@@ -3,7 +3,7 @@
   "release_state": "released",
   "source_state": "release",
   "installer_version": "0.20.0-rc.2",
-  "tests": 3456,
+  "tests": 3477,
   "redteam_faelle": 277,
   "kernel_zeilen": 996,
   "werkzeuge": 37,

site/vergleich/index.html

@@ -290,7 +290,7 @@ <h1 id="main-content" tabindex="-1">The others can do more.</h1>
     <div class="spec-row">
       <div class="spec"><b>37</b>tools here, every one gated</div>
       <div class="spec"><b>100+</b>skills there  ·  counted honestly</div>
-      <div class="spec"><b>3456</b>automated tests</div>
+      <div class="spec"><b>3477</b>automated tests</div>
       <div class="spec"><b>277/277</b>adversarial cases</div>
       <div class="spec"><b>996</b>gate lines</div>
       <div class="spec"><b>0</b>inbound ports</div>

talos/cli.py

@@ -52,6 +52,7 @@
   setup [terminal|identity|model|mail]    guided setup; preserves other settings
   doctor [--online]                  what is missing — changes nothing
   config list|get|set|describe|validate    find, explain and change settings
+  import list|hermes|openclaw           preview portable data from another agent
   models [--refresh]                 which models a provider offers
   local-check [--provider ollama|lm-studio]    probe local endpoint without operator data
 
@@ -266,6 +267,65 @@ def cmd_chat(rest: list[str], out=None) -> int:
     return 0
 
 
+def cmd_import(rest: list[str], out=None) -> int:
+    """Preview portable data from another agent without changing either home.
+
+    Import text is persistent, foreign instruction material.  The first release
+    therefore stops at a redacted plan: activation needs a separate reviewed path,
+    especially for skills.  Keeping this command before configuration loading also
+    lets a new Talos installation inspect its old home before setup.
+    """
+    schreiben = (out or sys.stdout).write
+    usage = "  usage: talos import <hermes|openclaw> --source <absolute-path> --dry-run\n"
+    if rest in (["--help"], ["-h"]):
+        schreiben(
+            usage
+            + "         talos import list\n"
+            + "  Preview only. Secrets, permissions, executable hooks, schedules and sessions are skipped.\n"
+        )
+        return 0
+    if rest == ["list"]:
+        schreiben(
+            "\n  hermes     portable profile, memory and Agent Skills\n"
+            "  openclaw   portable workspace profile, memory and Agent Skills\n\n"
+            "  Dry run only; imported text remains inactive and grants no permission.\n\n"
+        )
+        return 0
+    if not rest or rest[0] not in {"hermes", "openclaw"}:
+        schreiben(usage)
+        return 2
+    source_type = rest[0]
+    source = ""
+    dry_run = False
+    index = 1
+    while index < len(rest):
+        argument = rest[index]
+        if argument == "--source" and index + 1 < len(rest) and not source:
+            source = rest[index + 1]
+            index += 2
+            continue
+        if argument == "--dry-run" and not dry_run:
+            dry_run = True
+            index += 1
+            continue
+        schreiben(f"  unknown or repeated import option: {argument}\n{usage}")
+        return 2
+    if not source or not dry_run:
+        schreiben(usage + "  This release requires an explicit source and supports preview only.\n")
+        return 2
+
+    from .import_agents import ImportPlanError, plan_import
+    from .paths import PROFILE_DIR
+
+    try:
+        plan = plan_import(source, PROFILE_DIR, source_type=source_type)
+    except (ImportPlanError, OSError) as error:
+        schreiben(f"\n  import preview refused: {error}\n\n")
+        return 1
+    schreiben("\n" + plan.render() + "\n\n")
+    return 0
+
+
 def _help(_rest: list[str]) -> int:
     from .terminalui import heading, help_text
 
@@ -281,6 +341,7 @@ def _help(_rest: list[str]) -> int:
     "setup": _lazy("setup_wizard", "run_setup"),
     "doctor": _lazy("doctor", "run_doctor"),
     "config": _lazy("configcli", "run_config"),
+    "import": cmd_import,
     "models": _lazy("models", "run_models"),
     "local-check": _lazy("localcheck", "run_local_check"),
     "ask": cmd_ask,
@@ -337,4 +398,4 @@ def dispatch(args: list[str]) -> int | None:
 
 COMMANDS = tuple(TABLE)
 
-__all__ = ["COMMANDS", "HELP", "TABLE", "cmd_status", "cmd_verify", "cmd_version", "dispatch"]
+__all__ = ["COMMANDS", "HELP", "TABLE", "cmd_import", "cmd_status", "cmd_verify", "cmd_version", "dispatch"]