diff --git a/README.md b/README.md
index 0e8779f..a9c369a 100644
--- a/README.md
+++ b/README.md
@@ -57,7 +57,12 @@ Published images use these tags:
 
 - `sha-<12-character-commit>` for every push;
 - `latest` for `master`;
-- the exact `v*` tag for releases.
+- the `v*` release ref, with Docker-invalid characters replaced by `-`.
+
+The Gitea workflow reads `HARBOR_REGISTRY`, `HARBOR_USER`, and
+`HARBOR_PASSWORD`. `HARBOR_REGISTRY` may be a host such as `reg.rsxx.ru` or an
+HTTP(S) URL; the workflow strips the scheme and trailing slash before composing
+Docker image references. The Harbor project is appended separately as `rsmon`.
 
 ## systemd
 
@@ -113,9 +118,17 @@ has only `CAP_NET_RAW` for ICMP checks.
 | `WORKER_CLUSTER_PEERS` | no | none | Comma-separated `node@host:port` peers. |
 | `WORKER_CLUSTER_DATA_DIR` | with cluster | none | Persistent Raft state directory. |
 
-The public liveness endpoint is `GET /healthz`. `rsmon-worker health` checks the
-configured control plane's `/up` endpoint and is suitable for container health
-checks.
+The public liveness endpoint is `GET /healthz`; the image probes it with
+`rsmon-worker liveness`. `rsmon-worker health` separately checks the configured
+control plane's `/up` endpoint for connectivity diagnostics.
+
+## Implementation Documentation
+
+Worker architecture, protocol, inventory, private-worker isolation, network
+diagnostics, web console, and critical-cluster work are specified in
+[`docs/README.md`](docs/README.md). These documents replace worker-owned planning
+material formerly kept in the RSMon control-plane repository and include a
+source migration ledger.
 
 ## Security
 
diff --git a/docs/README.md b/docs/README.md
new file mode 100644
index 0000000..9907524
--- /dev/null
+++ b/docs/README.md
@@ -0,0 +1,72 @@
+# Worker Implementation Documents
+
+This directory is the implementation authority for the standalone RSMon worker.
+It converts the worker-related plans formerly kept in the RSMon control-plane
+repository into contracts tied to this repository's packages and tests.
+
+These documents distinguish three states:
+
+- **Implemented**: wired into `cmd/rsmon-worker` and covered by tests.
+- **Partial**: code exists, but a named runtime path or acceptance criterion is
+  missing.
+- **Planned**: no production path exists yet. The document identifies the
+  package, protocol, dependency, and tests required to implement it.
+
+Plans do not override code. When a document and the current implementation
+disagree, update both in the same change or mark the discrepancy explicitly.
+
+## Implementation Index
+
+| Document | Scope | Current state |
+| --- | --- | --- |
+| [architecture.md](architecture.md) | Process boundaries, ownership, runtime composition | Implemented with partial subsystems |
+| [control-plane-protocol.md](control-plane-protocol.md) | WebSocket frames, leases, config, compatibility APIs | Implemented |
+| [tasks-and-notifications.md](tasks-and-notifications.md) | Normal checks and delegated notification execution | Implemented with provider gaps |
+| [web-console-and-observability.md](web-console-and-observability.md) | Local UI, auth, inventory view, host telemetry | Partial |
+| [inventory.md](inventory.md) | Worker host discovery and control-plane inventory projection | Partial |
+| [private-workers.md](private-workers.md) | Customer worker isolation, credentials, onboarding | Partial |
+| [network-diagnostics.md](network-diagnostics.md) | Confirmation checks and dedicated diagnostic tasks | Partial |
+| [critical-check-cluster.md](critical-check-cluster.md) | Raft-backed dispatchless critical checks | Scaffold only |
+| [implementation-roadmap.md](implementation-roadmap.md) | Ordered repository work packages and release gates | Active |
+| [source-plan-migration.md](source-plan-migration.md) | Source-to-target conversion ledger and resolved conflicts | Complete mapping |
+
+## Ownership Boundary
+
+This repository owns:
+
+- worker process startup, configuration, and graceful shutdown;
+- the control-plane client and wire types;
+- check and notification execution;
+- local web console, local state, inventory collection, and host metrics;
+- worker-to-worker transport and Raft state;
+- Docker, Compose, systemd, and release-image packaging.
+
+The RSMon control-plane repository owns:
+
+- users, accounts, monitors, contacts, billing, and RBAC;
+- PostgreSQL task production, selection, leasing, retry, and dead-letter state;
+- worker registration and authorization policy;
+- accepted-result application and VictoriaMetrics persistence;
+- customer-facing fleet, inventory, diagnostics, and incident APIs.
+
+The worker must not receive control-plane PostgreSQL, Valkey, or
+VictoriaMetrics credentials. Normal results, host telemetry, inventory, and
+diagnostics are reported over authenticated protocol messages for validation
+and persistence by the control plane.
+
+## Required Quality Gates
+
+Run before merging worker changes:
+
+```bash
+make check
+docker build -t rsmon-worker:test .
+docker run --rm rsmon-worker:test --version
+docker compose config
+```
+
+Features that alter protocol messages require compatibility tests in
+`internal/wire` and runner protocol tests in `internal/distworker`. Raft work
+requires deterministic FSM tests and a multi-node fault test. Security-sensitive
+features require negative tests for account scope, arbitrary target rejection,
+credential redaction, and unauthenticated access.
diff --git a/docs/architecture.md b/docs/architecture.md
new file mode 100644
index 0000000..821db98
--- /dev/null
+++ b/docs/architecture.md
@@ -0,0 +1,85 @@
+# Worker Architecture
+
+## Runtime
+
+`cmd/rsmon-worker/main.go` composes three independently stoppable subsystems:
+
+1. `internal/distworker`: persistent control-plane connection, normal task
+   execution, selfcheck, notification delivery, and server metric reports.
+2. `internal/webapp`: authenticated local operator console and SQLite-backed
+   local state.
+3. `internal/workercluster`: optional worker-to-worker Raft transport and FSM.
+
+All three share one cancellation context. The worker can run without the web
+console using `--no-web`; the cluster is disabled unless
+`WORKER_CLUSTER_ENABLED=true`.
+
+## Two Execution Paths
+
+The two check paths are deliberately separate.
+
+| Path | Work origin | Ownership | Worker implementation |
+| --- | --- | --- | --- |
+| Normal checks and notifications | Control-plane PostgreSQL `tasks` rows | Exact lease token and expiry held by control plane | `internal/distworker` |
+| `distributed_critical` | Signed config adopted by a worker Raft cluster | Raft FSM incident and outbox metadata | `internal/workercluster` plus a planned scheduler bridge |
+
+Normal work always uses control-plane dispatch and leases. The critical path
+must never use a normal task row, job assignment, or lease. Adding the critical
+path must not change normal protocol behavior.
+
+## Package Boundaries
+
+| Package | Responsibility | Must not do |
+| --- | --- | --- |
+| `internal/wire` | JSON-compatible protocol types | Network or persistence work |
+| `internal/distworker` | Connections, queues, pools, execution orchestration | Direct control-plane database writes |
+| `internal/checkexec` | Database-free check dispatch | Scheduling or result persistence |
+| `checks/*` | Protocol-specific probes | Worker selection or account policy |
+| `internal/webapp` | Local UI, sessions, audit, local snapshots | Become the fleet source of truth |
+| `internal/workercluster` | Raft transport, log, FSM, snapshots, membership | Store raw samples or plaintext credentials |
+| `internal/sender` | Provider calls shared by worker executors | Own retries or durable task state |
+
+New host discovery should move into `internal/inventory`; the web console may
+read it through an interface but must not remain the collector's owner. New
+network probes should live in `internal/netdiag` and use an allowlisted target
+provided by a validated control-plane task.
+
+## Current Status
+
+| Capability | State | Evidence |
+| --- | --- | --- |
+| WebSocket normal check execution | Implemented | `internal/distworker/client.go`, `runner.go` |
+| Check kinds | Implemented | `internal/checkexec/exec.go`, `checks/*` |
+| Delegated notifications | Partial | email, Telegram, webhook, Mattermost implemented; SMS/voice unsupported |
+| Host telemetry to control plane | Implemented | `internal/distworker/server_metrics.go` |
+| Local web console | Implemented | `internal/webapp/server.go`, `routes.go` |
+| Local inventory and status pages | Partial | collectors exist but are not started by `Server.Start` |
+| Private-worker scheduler isolation | Implemented in shared selector code | `app/models/worker_node.go`, `task_selector.go`, `check_jobs.go` |
+| Private-worker bootstrap and worker-side scope verification | Planned | no one-time bootstrap or signed config path |
+| Cross-worker confirmation | Partial | peer selfcheck and shared model logic exist; no rich diagnostic tasks |
+| Raft membership and persistence | Partial | `internal/workercluster` |
+| Dispatchless critical checks | Planned | no scheduler, signature adoption, quorum incident engine, or outbox executor |
+
+## Data Rules
+
+- Credentials delivered in `init`/`config` stay in memory and are never written
+  to the local SQLite store, Raft log, or snapshots.
+- Raw check samples and high-volume telemetry never enter Raft.
+- Local UI state is not authoritative control-plane state.
+- A normal result is not valid without the current exact lease token.
+- Any worker-originated account, server, monitor, or target identifier is
+  untrusted until the control plane validates it against the authenticated
+  worker.
+- A worker must not execute a synthetic health check whose target is itself.
+
+## Runtime Completion Criteria
+
+- SIGTERM drains the runner, stops collectors, closes listeners, shuts down
+  Raft, and closes SQLite without exceeding the service stop timeout.
+- A web-console failure is reported and causes an intentional process policy;
+  it must not silently leave a partially healthy process.
+- Startup rejects malformed `WORKER_URL`, incomplete auth credentials, invalid
+  cluster settings, and unwritable data directories before accepting work.
+- `GET /healthz` reports process-local liveness. Control-plane reachability is
+  a separate readiness/selfcheck signal and must not make a healthy container
+  fail its local liveness probe.
diff --git a/docs/control-plane-protocol.md b/docs/control-plane-protocol.md
new file mode 100644
index 0000000..bfb7efe
--- /dev/null
+++ b/docs/control-plane-protocol.md
@@ -0,0 +1,156 @@
+# Control-Plane Protocol
+
+## Authority
+
+`internal/wire/types.go` is the executable schema. New fields must be optional
+during rollout, and new branches require decode tests against both old and new
+frames. Version strings are informational; explicit capabilities select
+protocol features.
+
+The worker connects to:
+
+```text
+GET /worker?token=<RSMON_TOKEN>
+```
+
+`/api/worker` and the HTTP jobs/results APIs remain compatibility paths. New
+workers use WebSocket task envelopes.
+
+## Frame Model
+
+Every frame is a `wire.WorkerMessage` with `kind` and one active content
+branch.
+
+| Direction | `kind` | Active branch | State |
+| --- | --- | --- | --- |
+| Server to worker | `init` or `config` | `init` | Implemented |
+| Server to worker | `task` | `task_envelope` | Implemented |
+| Worker to server | `result` | `result` | Implemented |
+| Worker to server | `result` | `notification_result` | Implemented |
+| Worker to server | `result` | `server_metric` | Implemented |
+| Worker to server | `heartbeat` | `heartbeat` | Implemented |
+| Either | `error` | `error` | Implemented |
+| Worker to server | `result` | inventory report | Planned |
+| Server to worker | cluster config/witness | dedicated typed branch | Planned |
+| Server to worker | diagnostic task | new task-envelope variant | Planned |
+
+The legacy top-level `task` and `notification_task` branches are accepted for
+rollout compatibility. If a frame includes both a valid current
+`task_envelope` and a legacy branch, the worker executes only
+`task_envelope`. The current runner ignores malformed or unsupported task
+branches and does not yet compare the envelope job ID with the inner job ID;
+strict rejection and reporting are P1 work below.
+
+## Initialization And Refresh
+
+`wire.WorkerInit` supplies runtime values owned by the control plane:
+
+- worker ID, region, advertised URL, capabilities, and concurrency;
+- allowed notification methods and account IDs;
+- optional linked server ID for host metrics;
+- LLM endpoints;
+- scoped notification credentials and system contacts;
+- peer workers for selfcheck/cluster-adjacent behavior.
+
+The worker clamps supplied concurrency to its local maximum. Credentials are
+replaced atomically in memory on refresh. Removed credentials must become
+unavailable immediately after the refresh is applied.
+
+Private-worker hardening will add an immutable worker account ID, config
+version, expiry, and signature. Until then the executable trusts the
+authenticated control plane to send a correctly scoped config; server-side
+selection remains the primary isolation boundary.
+
+## Normal Task Envelope
+
+`wire.TaskEnvelope` is a tagged union:
+
+```json
+{
+  "kind": "task",
+  "task_envelope": {
+    "type": "check",
+    "job_id": "uuid",
+    "check": {
+      "job_id": "uuid",
+      "lease_token": "per-lease-secret",
+      "check_id": 123,
+      "monitor_id": 456,
+      "kind": "http",
+      "host": "example.com",
+      "url": "https://example.com",
+      "interval": 60,
+      "settings": {}
+    }
+  }
+}
+```
+
+`type=notification` activates `notification` instead. Job IDs in the envelope
+and active branch must match once P1 validation lands. Today a task without a
+recognized populated branch is ignored without execution.
+
+## Result Invariants
+
+- Echo `job_id` and the exact `lease_token` from the task.
+- Send one terminal result per execution attempt.
+- Never retry a result by executing the task again. Result transport retries
+  resend the same terminal report.
+- Treat duplicate terminal acknowledgements as success.
+- Do not infer task acceptance from a WebSocket write alone; durable ownership
+  remains on the control plane until it validates the result.
+- Bound error strings and provider responses before transmission.
+
+## Host Metrics
+
+`wire.ServerMetricReport` is sent only after `WorkerInit.ServerID` is present.
+The worker collects locally and sends bounded snapshots. The control plane
+validates worker/server/account ownership and persists both the latest cache
+and VictoriaMetrics points. The worker does not have TSDB credentials.
+
+## Protocol Work Packages
+
+### P1: Conformance Tests
+
+Files:
+
+- `internal/wire/types_test.go`
+- `internal/distworker/runner_protocol_test.go`
+
+Tests:
+
+- decode every current frame branch;
+- prefer the current envelope over duplicate legacy branches;
+- reject mismatched envelope and inner job IDs;
+- preserve unknown optional fields during compatible rollout;
+- reject missing lease tokens before execution result submission;
+- atomically replace config and credentials.
+
+### P2: Safe Token Rotation
+
+Current `Runner.RotateToken` closes the runner stop channel. Replace this with a
+connection-scoped cancellation path that stores the new token, closes only the
+active WebSocket, and reconnects without terminating worker services.
+
+Acceptance: rotating from the web console produces a reconnect using the new
+token while the HTTP listener, collectors, and optional cluster stay running.
+
+### P3: Signed Private-Worker Config
+
+Add to the init/config branch:
+
+- `account_id` for private workers;
+- monotonic `config_version`;
+- `issued_at` and `expires_at`;
+- signature key ID and Ed25519 signature over canonical payload bytes.
+
+Reject regressions, invalid signatures, expired config, and account changes.
+Keep the last valid config only until its expiry; do not silently accept an
+invalid replacement.
+
+### P4: New Typed Branches
+
+Inventory, diagnostic, and critical-cluster messages each receive a dedicated
+wire type. Do not tunnel them through `event` or arbitrary `json.RawMessage`.
+Each branch must define payload limits, account/target validation ownership,
+idempotency, and compatibility behavior before implementation.
diff --git a/docs/critical-check-cluster.md b/docs/critical-check-cluster.md
new file mode 100644
index 0000000..dcd8ae5
--- /dev/null
+++ b/docs/critical-check-cluster.md
@@ -0,0 +1,188 @@
+# Dispatchless Critical-Check Cluster
+
+## Status
+
+`internal/workercluster` is a scaffold, not a production critical-check engine.
+It currently provides Hashicorp Raft, bbolt log/stable storage, snapshots,
+HTTP-based authenticated transport, bootstrap/join/membership operations, log
+entry types, a small FSM, and operator status.
+
+It does not yet execute `distributed_critical` checks, verify signed config,
+evaluate observation/region/notification quorum, encrypt snapshots, deliver a
+commit-backed outbox, or consume an external witness report.
+
+The hardcoded test-config endpoint and startup flag are development-only and
+must stay disabled in production.
+
+## Non-Negotiable Separation
+
+`distributed_critical` is a distinct path:
+
+- no control-plane job dispatch;
+- no PostgreSQL task row;
+- no task lease or single worker owner;
+- every eligible observer executes the adopted check deterministically;
+- raw samples go to the control plane/TSDB outside Raft;
+- classified observations, compact incidents, and outbox metadata enter Raft;
+- normal checks and normal notifications continue unchanged.
+
+The control plane signs configuration, receives replay, and acts as an external
+witness. It is never a Raft voter.
+
+## Topology
+
+- Production clusters have 3 or 5 voters, never an even count.
+- Voters remain within a bounded-latency topology. Remote regions use observer
+  nodes unless measured RTT supports the configured election timeout.
+- Nodes may be `voter`, `observer`, or `voter+observer`.
+- Observer count affects observation policy, not Raft election quorum.
+- A node without durable, real-fsync storage cannot be a voter.
+
+Single-voter bootstrap is temporary. The cluster must not execute customer
+critical checks until at least three voters are healthy and the signed
+observer set is committed.
+
+## Raft State
+
+The FSM contains only:
+
+- monotonic config version and adopted critical-check definitions;
+- versioned observer set;
+- compact incident state and committed observations;
+- bounded notification outbox metadata and idempotency tombstones;
+- compact member diagnostics and partition/witness state;
+- tenant isolation policy and applied-version index.
+
+It never contains raw probe samples, response bodies, provider bodies,
+plaintext credentials, credential envelopes, or normal task leases.
+
+`raft-boltdb/v2` is only the local Raft `LogStore` and `StableStore`; it is not
+the application model. Snapshots are produced by the FSM and must be versioned,
+checksummed, and encrypted at rest before production use.
+
+## Log Entries
+
+Allowed application entry kinds:
+
+- `config.adopt`
+- `observer_set.update`
+- `incident.observe`
+- `incident.transition`
+- `outbox.enqueue`
+- `outbox.delivered`
+- `outbox.ack`
+- `partition.report`
+- `membership.propose_add`
+- `membership.demote`
+- `membership.remove`
+- `diagnostics.update`
+
+No `task.lease.*`, retry, completion, or dead-letter entries belong in this
+package.
+
+## Signed Config Adoption
+
+The control plane sends an Ed25519-signed canonical payload containing:
+
+- cluster and tenant IDs;
+- monotonic config version and expiry;
+- critical checks and incident policies;
+- proposed observer set, role/region map, and content hash;
+- signing key ID and credential-envelope references.
+
+Every node verifies signature, cluster identity, expiry, monotonic version,
+observer-set hash, and supported schema. The leader proposes `config.adopt` and
+`observer_set.update`. No observer executes the new config until both commits
+are applied locally. Invalid config is rejected and reported through compact
+diagnostics; there is no silent downgrade.
+
+## Scheduling And Observation
+
+For each adopted check, eligible observers derive the same interval boundary
+from `epoch + n*interval`. Per-worker jitter is deterministic from worker ID,
+check ID, observer-set version, and interval number. At each tick:
+
+1. Run the underlying probe through `internal/checkexec` outside Raft.
+2. Send raw metrics through the normal control-plane metrics path.
+3. Classify to `ok`, `warn`, `down`, or `unknown` using signed policy.
+4. Propose `incident.observe` to the leader with committed timestamp and
+   observer/config versions.
+
+Followers forward proposals or return a typed not-leader response containing
+the current leader. They never silently drop observations.
+
+## Four Quorums
+
+- `raft_quorum`: majority of voters required to commit.
+- `observation_quorum`: observers agreeing within the observation window.
+- `region_quorum`: represented regions agreeing on state.
+- `notification_quorum`: region agreement required for first customer alert.
+
+All names remain distinct in code, metrics, configuration, and logs.
+
+## Deterministic Incident FSM
+
+`FSM.Apply` must perform no I/O, network calls, wall-clock reads, randomness,
+or provider calls. It uses only committed payload values and prior state.
+
+Incident lifecycle is `clear -> observing -> open -> resolving -> clear`.
+Policy includes confirmation count, observation window, classification,
+minimum incident dwell, transition-rate suppression, cooldown, and outbox retry
+metadata. Apply enforces bounded state and deterministic idempotency keys.
+
+The current FSM merely records the latest observation and accepts externally
+constructed transitions. Replace this placeholder with deterministic policy
+evaluation and tests before running real checks.
+
+## Commit-Before-Notify
+
+Strict order:
+
+1. Probe and classify outside Raft.
+2. Commit observation at `raft_quorum`.
+3. FSM deterministically commits incident transition and outbox metadata.
+4. After the committed outbox entry is visible, an executor calls the provider
+   outside Raft.
+5. Provider acknowledgement produces `outbox.delivered`; retry/failure produces
+   bounded metadata updates.
+
+Provider credentials stay in node-local memory/secure storage and are never
+part of an entry or snapshot. Delivery uses stable channel idempotency keys to
+survive leader failover.
+
+## Partitions And Witness
+
+Partition states are `steady`, `degraded`, `partitioned`, `healing`,
+`split_brain_detected`, and `witness_only`. Loss of `raft_quorum` freezes
+incident transitions and outbox creation. A cluster must never fabricate an
+open or recovery while partitioned.
+
+The external witness reports observed leader/term, reachable voters and
+observers, split-brain flag/time, and report time. It can alert and inform state
+but cannot commit or fabricate incidents.
+
+## Ordered Implementation
+
+1. Remove production exposure of debug config application.
+2. Complete versioned FSM types, command validation, and deterministic tests.
+3. Add mTLS identity and safe one-claim cluster bootstrap.
+4. Add signed config and observer-set adoption.
+5. Add deterministic scheduler and checkexec bridge in shadow mode.
+6. Implement observation aggregation, incident policy, and idempotency.
+7. Implement encrypted snapshots and restore/migration tests.
+8. Add metadata outbox executor and failover-safe delivery.
+9. Add external witness, replay, metrics, and operational runbooks.
+10. Run synthetic 3/5-node fault campaigns before any customer check.
+
+## Release Gates
+
+- Two independent bootstraps for one cluster are detected and rejected.
+- Changes that leave fewer than three or an even number of voters are rejected.
+- Leader failover completes within twice the configured election timeout.
+- A minority partition commits zero incident transitions.
+- Loss of observation quorum with intact Raft quorum sends no notification.
+- Snapshot inspection finds no raw samples or credential material.
+- Signature mutation and observer-set version mismatch prevent execution.
+- Killing the leader between provider acknowledgement and metadata commit does
+  not create duplicate customer-visible notification effects.
+- Normal WebSocket tasks and notifications continue throughout cluster tests.
diff --git a/docs/implementation-roadmap.md b/docs/implementation-roadmap.md
new file mode 100644
index 0000000..f9995b9
--- /dev/null
+++ b/docs/implementation-roadmap.md
@@ -0,0 +1,165 @@
+# Worker Implementation Roadmap
+
+This roadmap orders work by dependency and risk. A package existing does not
+make a work package complete; its release gate must pass in a wired process.
+
+## R0: Distribution Reliability
+
+State: active.
+
+Worker repository:
+
+- publish valid Harbor references after normalizing a configured registry URL;
+- use local `/healthz` for container liveness;
+- verify amd64 and arm64 image startup, Chromium availability, and version
+  metadata;
+- add a package/install smoke test for Docker and systemd artifacts;
+- document immutable SHA and release tags as production defaults.
+
+Gate: a push publishes `sha-<12>` and `latest` manifests for both platforms,
+and a container remains healthy when the control plane is unavailable.
+
+## R1: Runtime Correctness
+
+State: partial.
+
+Worker repository:
+
+- wire local inventory and metrics collector lifecycle into web server start
+  and shutdown;
+- validate full HTTP config, including `WORKER_URL`, in main startup;
+- make token rotation reconnect rather than stop the runner;
+- define process policy when the web listener exits unexpectedly;
+- add WebSocket reconnect, drain, duplicate frame, and backpressure tests.
+
+Gate: collectors populate real pages, rotation preserves all subsystems, and
+SIGTERM leaves no listener, collector, task, or SQLite goroutine behind.
+
+## R2: Protocol And Credential Hardening
+
+State: partial.
+
+Worker repository:
+
+- complete frame conformance tests;
+- add signed config version/account identity;
+- enforce task account and credential scope locally;
+- add bounded delegated-delivery audit to the local console;
+- support per-account webhook/Mattermost credentials after control-plane wire
+  support exists.
+
+Control-plane dependency:
+
+- signed config producer and key rotation;
+- persisted account-scoped webhook/Mattermost credentials;
+- lease/task account fields treated as immutable during result application.
+
+Gate: cross-account fixtures fail before network execution and no credential
+appears in worker logs, SQLite, snapshots, or protocol error payloads.
+
+## R3: Private-Worker Lifecycle
+
+State: planned beyond scheduler isolation.
+
+Worker repository:
+
+- one-time bootstrap exchange and atomic secret storage;
+- token handoff and revocation handling;
+- stale/expired signed-config behavior;
+- clear disabled/revoked UI state;
+- optional mTLS client identity.
+
+Control-plane dependency:
+
+- customer worker CRUD and billing entitlement;
+- one-time bootstrap token state;
+- immutable account binding and signed config;
+- audit and revocation acknowledgement.
+
+Gate: a customer can install, connect, execute only their work, rotate, disable,
+revoke, and uninstall without an operator admin secret.
+
+## R4: Host Inventory And Console
+
+State: partial.
+
+Worker repository:
+
+- extract `internal/inventory` and start process/host collection;
+- add typed bounded inventory report;
+- add Docker/Compose, nginx, systemd, and address collectors in stages;
+- add read-only Compose status/logs before mutation;
+- correct local interval metrics and add optional hardware sections.
+
+Control-plane dependency:
+
+- worker inventory report ingestion and source-aware reconciliation;
+- server/worker/account ownership validation;
+- deploymentd comparison and monitor suggestion UI.
+
+Gate: partial failures cannot delete inventory, worker and deploymentd sources
+coexist, and reports contain no environment or credential values.
+
+## R5: Dedicated Network Diagnostics
+
+State: planned; normal confirmation is already partial.
+
+Worker repository:
+
+- add diagnostic wire branch and `internal/netdiag`;
+- ship HTTP GET/HEAD and unauthenticated SSH handshake first;
+- add TCP/DNS, then separately gate traceroute/MTR;
+- implement SSRF, redirect, timeout, output, and rate protections.
+
+Control-plane dependency:
+
+- monitor-bound target normalization and signed target hash;
+- diagnostic task production, storage, RBAC, rate limits, and UI;
+- private-worker source eligibility.
+
+Gate: no task can probe outside its authorized monitor target and all results
+are bounded, structured, and credential-free.
+
+## R6: Public Check Economy
+
+State: planned.
+
+Depends on R2, R3, and diagnostic-grade target protection. Public work uses a
+separate signed grant and safe HTTP methods, never widened account scope.
+
+Gate: private targets and credential-bearing requests are impossible under DNS
+rebinding and redirects; accounting remains idempotent under result replay.
+
+## R7: Critical-Check Cluster
+
+State: Raft scaffold only.
+
+Worker repository:
+
+- secure bootstrap and mTLS identity;
+- signed config adoption and observer-set versioning;
+- deterministic scheduler and check executor bridge;
+- deterministic incident/quorum FSM;
+- encrypted snapshots and restore;
+- metadata outbox, witness, replay, and metrics;
+- 3/5-node fault campaigns.
+
+Control-plane dependency:
+
+- critical-check config/signing service;
+- credential-envelope service;
+- witness and replay endpoints;
+- customer incident projection and audit.
+
+Gate: all release gates in
+[critical-check-cluster.md](critical-check-cluster.md) pass while normal tasks
+continue without behavioral change.
+
+## Documentation Rule
+
+Every completed work package updates:
+
+- the relevant implementation document's current-state section;
+- this roadmap state and gate evidence;
+- root configuration examples when environment or deployment changes;
+- the source migration ledger if a control-plane plan is superseded.
diff --git a/docs/inventory.md b/docs/inventory.md
new file mode 100644
index 0000000..485f801
--- /dev/null
+++ b/docs/inventory.md
@@ -0,0 +1,165 @@
+# Worker Host Inventory
+
+## Purpose And Boundary
+
+Inventory connects checks to the host, service, Compose project, domain, and
+deployment that they observe. The worker collects facts; the RSMon control
+plane validates identity, applies account scope, reconciles lifecycle, and
+persists the inventory projection.
+
+The control-plane inventory entities remain `Server`, `ServerIp`, `Site`,
+`Deployment`, `Domain`, `Repo`, and `site_repos`. Their API and rstuff stream
+projection are control-plane concerns. This repository owns only local
+discovery and worker-originated reports.
+
+## Existing Collectors
+
+Current state:
+
+- `internal/webapp/inventory.go` discovers Linux processes, command lines,
+  working directories, TCP listeners, and basic resource facts.
+- snapshots are stored in the local web-console SQLite database;
+- `internal/distworker/server_metrics.go` reports bounded process/network host
+  snapshots, but not normalized application inventory;
+- external `deploymentd` remains the implemented authoritative collector for
+  nginx, Docker Compose, and host inventory on the control plane.
+
+The worker does not currently send an inventory report to the control plane.
+It must not claim deploymentd parity until the protocol, ingestion, and
+reconciliation tests below are complete.
+
+## Target Package Layout
+
+Move collection ownership out of `internal/webapp`:
+
+```text
+internal/inventory/
+  collector.go       orchestration and partial-success envelope
+  process_linux.go   procfs process and listener discovery
+  docker.go          Docker and Compose discovery behind capability
+  nginx.go           read-only nginx virtual-host discovery
+  systemd.go         allowlisted unit discovery
+  normalize.go       stable IDs and control-plane report conversion
+  store.go           optional local snapshot interface
+```
+
+The web console consumes a read-only snapshot interface. The runner consumes a
+bounded report interface. Collectors return per-section errors; one failed
+section does not discard successful sections.
+
+## Report Contract
+
+Add a typed `WorkerInventoryReport` branch to `internal/wire`:
+
+```json
+{
+  "kind": "result",
+  "inventory": {
+    "schema": "rsmon.worker.inventory.v1",
+    "report_id": "uuid",
+    "server_id": 42,
+    "collected_at": "2026-07-13T15:00:00Z",
+    "full_snapshot": true,
+    "host": {},
+    "addresses": [],
+    "processes": [],
+    "compose_projects": [],
+    "nginx_sites": [],
+    "systemd_units": [],
+    "errors": [{"section": "docker", "code": "unavailable"}]
+  }
+}
+```
+
+Constraints:
+
+- `report_id` is stable across transport retries and unique across collection
+  runs.
+- `server_id` comes from authenticated `WorkerInit.ServerID`, never local env.
+- A full snapshot permits missing-item reconciliation; partial reports do not.
+- Section and total item counts, strings, labels, and serialized bytes are
+  bounded before enqueue.
+- Environment variable values, file contents, process environments, container
+  secrets, and Compose `.env` values are never included.
+- The control plane rejects a report when the worker is not linked to the
+  server or the private worker's account does not own it.
+
+## Normalization
+
+Stable local identity keys:
+
+- process app: executable plus canonical working directory;
+- Compose site: Docker Compose project name plus canonical project directory;
+- Compose deployment: project plus service name;
+- nginx deployment: canonical config path;
+- domain: normalized lowercase hostname without trailing dot;
+- systemd service: unit name.
+
+Container IDs and PIDs are observations, not stable entity IDs. Never use them
+as the sole upsert key. Every deployment includes `last_seen_at`; only a full
+successful section can mark previously observed entities missing.
+
+## Collector Requirements
+
+### Processes
+
+- Group related processes deterministically by executable, working directory,
+  and parent relationship.
+- Parse TCP and UDP listeners for IPv4 and IPv6.
+- Do not read another process's environment.
+- Exclude kernel threads and the worker itself from application suggestions.
+
+### Docker And Compose
+
+- Disabled without an explicit Docker capability.
+- Prefer Docker/Compose JSON output and labels; do not parse human tables.
+- Collect project, service, image, state, ports, mount paths, health, and
+  Traefik host rules. Exclude environment values and registry credentials.
+- Treat Docker socket access as root-equivalent and report the capability to
+  the operator.
+
+### Nginx
+
+- Read only configured allowlisted roots.
+- Collect config path, listen values, server names, auth presence, root path,
+  and proxy presence. Do not send certificate private-key paths or contents.
+
+### Host And Server IPs
+
+- Collect hostname, OS/kernel summary, and valid non-loopback addresses.
+- Do not guess a primary address; report interface and route metadata so the
+  control plane can apply policy.
+
+## Relationship To deploymentd And rstuff
+
+Worker reports and deploymentd ingest are complementary producers of the same
+control-plane projection. During migration:
+
+- deploymentd remains authoritative for sections it currently reports;
+- worker inventory is feature-gated per server;
+- every row records source and source report ID;
+- reconciliation occurs per source and section to prevent one producer from
+  deleting another producer's observations;
+- the rstuff `rsmon.inventory.v1` stream remains a control-plane output/input,
+  not a worker protocol.
+
+## Delivery Phases
+
+1. Start and stop the existing process collector; move it behind an interface.
+2. Add typed report and control-plane ingestion for process/host/address data.
+3. Add read-only Docker and Compose discovery.
+4. Add nginx and systemd discovery.
+5. Add source-aware reconciliation and deploymentd comparison mode.
+6. Enable monitor suggestions only after operators confirm discovered targets.
+
+## Acceptance Tests
+
+- A worker restart preserves no false stable identity based on PID/container ID.
+- A failed Docker section does not erase successful nginx or process data.
+- Partial reports never mark absent deployments missing.
+- Full reports reconcile only rows owned by the same source and section.
+- Cross-account and worker/server mismatch reports are rejected.
+- Fixture snapshots prove no environment values or credential-like fields are
+  serialized.
+- Worker and deploymentd fixtures normalize equivalent Compose/nginx entities
+  to the same control-plane identity without deleting each other's rows.
diff --git a/docs/network-diagnostics.md b/docs/network-diagnostics.md
new file mode 100644
index 0000000..7b10d33
--- /dev/null
+++ b/docs/network-diagnostics.md
@@ -0,0 +1,127 @@
+# Network Diagnostics
+
+## Current Capability
+
+Two different features must not be conflated:
+
+1. **Cross-worker confirmation** is partially implemented. The control plane
+   can assign the same failed check to another operated worker, and peer
+   selfcheck/quorum helpers exist in `internal/distworker/peer.go`,
+   `consensus.go`, and `selfcheck.go`.
+2. **Rich diagnostic tasks** (`diag_http`, `diag_ssh`, DNS, TCP, traceroute,
+   MTR) are not implemented as worker protocol kinds.
+
+Confirmation remains a normal centrally leased check and does not require a
+new executor. Rich diagnostics require typed tasks and structured timing/path
+results.
+
+## Security Boundary
+
+Diagnostics can become a network scanner. Every diagnostic task must be bound
+to a monitor target already authorized and normalized by the control plane.
+The worker receives the normalized target plus a signed target hash and refuses
+runtime overrides.
+
+Required restrictions:
+
+- private worker account must equal task account;
+- source worker must not equal a target worker represented by the monitor;
+- no arbitrary host, URL, resolver, port range, or shell arguments;
+- resolve every hostname locally and reject prohibited IPs before connecting;
+- revalidate every HTTP redirect target;
+- enforce per-task timeout, output limit, concurrency, and destination rate;
+- never return authorization headers, request bodies, response bodies, SSH
+  credentials, or environment data;
+- audit allow and deny decisions on the control plane.
+
+## Target Package
+
+```text
+internal/netdiag/
+  types.go
+  validate.go
+  http.go
+  ssh.go
+  tcp.go
+  dns.go
+  traceroute_linux.go
+  mtr.go
+```
+
+`internal/netdiag` accepts a validated immutable request and returns a bounded
+result. It does not select workers, query monitors, persist data, or send
+protocol frames. `internal/distworker` owns task dispatch and result transport.
+
+## Protocol
+
+Extend `wire.TaskEnvelope` with `type=diagnostic` and a dedicated diagnostic
+branch. Do not encode diagnostics as check settings or untyped events.
+
+Common request fields:
+
+- job ID and exact lease token;
+- account, monitor, source worker, optional target worker, and region IDs;
+- kind and scheduling reason (`periodic`, `on_demand`, `failure_triggered`);
+- normalized target and target hash;
+- timeout and kind-specific bounded options.
+
+Common result fields:
+
+- job ID, lease token, kind, source and target IDs;
+- start/finish time and total latency;
+- `ok` or `err` status;
+- normalized error class;
+- bounded kind-specific data.
+
+Error classes are `timeout`, `dns_error`, `tcp_refused`, `tcp_reset`,
+`tls_handshake`, `tls_cert`, `http_status`, `auth_required`, `auth_failed`,
+`protocol_error`, `target_denied`, `unsupported_kind`, and `internal_error`.
+
+## Delivery Phases
+
+### D1: HTTP Diagnostics
+
+Implement GET/HEAD only. Capture DNS, connect, TLS, first-byte, total timing,
+status, final URL, bounded redirect chain, selected response headers, response
+size, TLS version/cipher, and certificate fingerprints. Do not return body
+content. Clamp total timeout to 30 seconds.
+
+### D2: SSH Diagnostics
+
+Perform banner, key exchange, and host-key inspection without authentication.
+Return DNS/connect/KEX timing, banner, selected algorithms, and SHA-256 host-key
+fingerprint. Password and private-key authentication are not part of v1.
+
+### D3: TCP And DNS
+
+TCP performs one destination connect and distinguishes timeout, refused, and
+reset. DNS resolves only the monitor hostname with the system or explicitly
+allowlisted resolver and returns bounded A/AAAA/CNAME answers and timing.
+
+### D4: Traceroute And MTR
+
+Linux-only, capability-gated, and disabled for private workers until abuse
+review. Use a native bounded implementation or a fixed executable with fixed
+arguments; never pass user strings to a shell. Return a maximum hop count and
+bounded probes per hop.
+
+## Relationship To Raft
+
+Normal diagnostic tasks remain centrally leased. Do not replicate raw
+diagnostic results in Raft. If duplicate failure-triggered diagnostics become a
+measured problem, the critical cluster may replicate only compact dedup
+metadata; this is not a prerequisite for diagnostic v1.
+
+## Acceptance Tests
+
+- Existing confirmation excludes the original failing worker and handles
+  duplicate results idempotently.
+- Diagnostic tasks cannot target a host or port different from their signed
+  monitor target.
+- HTTP redirects to prohibited addresses are rejected.
+- Private workers reject cross-account tasks before opening a socket.
+- HTTP and SSH fixture servers produce deterministic phase timings and bounded
+  metadata.
+- Timeouts cancel DNS/connect/TLS/read work and leave no goroutine behind.
+- Unsupported kinds return a structured terminal result.
+- Result serialization contains no credentials or response body.
diff --git a/docs/private-workers.md b/docs/private-workers.md
new file mode 100644
index 0000000..c63c704
--- /dev/null
+++ b/docs/private-workers.md
@@ -0,0 +1,147 @@
+# Private Workers
+
+## Model
+
+Operated and customer-operated workers run the same binary and protocol. The
+control plane assigns trust and scope. A private worker is untrusted for every
+account except its own and cannot widen its own scope through capabilities or
+protocol fields.
+
+Current shared scheduler code already supports restrictive account selection:
+
+- `app/models/worker_node.go` derives allowed check and notification accounts;
+- `app/models/check_jobs.go` filters check leasing;
+- `app/models/task_selector.go` filters generic task leasing;
+- a nonempty account list is restrictive; an empty list means platform worker
+  and is never available to a customer worker.
+
+This is only one isolation layer. The executable still trusts the authenticated
+control plane's init/task payload and private-worker onboarding is incomplete.
+
+## Required Invariants
+
+- A private worker has one immutable `account_id` in signed runtime config.
+- Every normal task includes `account_id`; the worker rejects mismatch before
+  execution and reports a protocol error without touching the target.
+- Credentials are filtered by account and method on the control plane, then
+  checked again by exact credential ID in the worker.
+- Private workers do not receive another account's monitor, contact, LLM, peer,
+  inventory, or notification data.
+- A private worker cannot become operated by changing local config or reported
+  capabilities.
+- Results cannot choose their account. The control plane applies them against
+  the leased task and authenticated worker.
+- Local UI data stays on the worker unless a typed, bounded protocol explicitly
+  permits upload.
+
+## Onboarding
+
+The current production path uses a worker token created by an administrator.
+The customer self-service target is:
+
+1. Control plane creates a disabled private-worker record and a one-time,
+   15-minute bootstrap token.
+2. Installer writes control-plane URL and bootstrap token to a mode-0600
+   temporary environment file.
+3. Worker exchanges it over TLS for a long-lived worker token and signed
+   immutable account identity.
+4. Worker writes the long-lived token to the configured secret file with mode
+   0600, removes the bootstrap token, and reconnects.
+5. Control plane activates the worker only after the first authenticated
+   heartbeat and valid capability report.
+
+The worker must never receive the control-plane registration admin secret.
+Bootstrap replay, expiry, worker-ID mismatch, or account mismatch fails closed.
+
+## Token Rotation
+
+Rotation is a two-token handoff:
+
+1. Control plane issues a replacement token and keeps the old token valid for a
+   short bounded overlap.
+2. Worker atomically writes the replacement secret.
+3. Worker closes only the active control-plane connection and reconnects.
+4. Successful authentication acknowledges rotation; control plane revokes the
+   old token.
+5. Failure retains the old token until overlap expires and emits an operator
+   warning.
+
+The current `Runner.RotateToken` terminates the runner by closing its stop
+channel. This must be fixed before claiming unattended rotation.
+
+## Runtime Config Authentication
+
+Private-worker init/config payloads add:
+
+- worker ID and immutable account ID;
+- monotonic config version;
+- issue and expiry timestamps;
+- allowed check kinds, notification methods, and public-task policy;
+- credential-set hash;
+- Ed25519 signature and key ID.
+
+The worker pins the control-plane verification key at bootstrap. It rejects
+signature failure, downgrade, expiry, account change, and unknown critical
+fields. Credentials remain in memory and are cleared when their signed scope
+expires.
+
+## Public Checks
+
+Cross-account public-check execution is not the same as normal private scope.
+It requires a separately signed `public_tasks` grant with:
+
+- explicit opt-in by worker owner;
+- allowed methods limited to safe public probes, initially HTTP GET/HEAD;
+- no credentials, custom authorization headers, request bodies, private IP
+  targets, or notification tasks;
+- independent concurrency and rate limits;
+- SSRF validation after every DNS resolution and redirect;
+- auditable reward/usage identity owned by the control plane.
+
+Do not represent public permission by making the private worker's account list
+empty or adding arbitrary account IDs.
+
+## Deployment Modes
+
+The systemd and Compose packages are both supported. Capabilities, not install
+type, control host access:
+
+- base: normal network checks and delegated notifications;
+- host metrics: procfs and statfs access;
+- inventory: read-only process/system discovery;
+- Docker discovery: explicit socket access, treated as root-equivalent;
+- Compose mutation: separate high-risk capability, disabled by default;
+- Raft voter: durable fsync-capable cluster data directory and mTLS transport.
+
+`CAP_NET_RAW` is granted only for ping/traceroute features. The worker remains
+unprivileged otherwise.
+
+## Worker Self-Monitoring
+
+A worker must not run a monitor that represents its own process or host. The
+control plane excludes it during selection; the worker also rejects a task that
+names its own worker/host identity once those fields are signed into the task.
+Peer health checks and control-plane selfcheck are separate from customer
+monitor execution.
+
+## Implementation Work Packages
+
+1. Add signed account/config identity to `internal/wire` and runner state.
+2. Validate task account and credential scope locally before dispatch.
+3. Implement one-time bootstrap and atomic token storage/rotation.
+4. Add worker disable/revoke behavior and visible stale-config state.
+5. Add mTLS as an optional first transport, then require it for Raft clusters.
+6. Add public-task grant and SSRF-safe executor only after private isolation is
+   proven.
+
+## Acceptance Tests
+
+- A private worker never leases or executes another account's normal task.
+- A forged capability or account field cannot widen scope.
+- Invalid, expired, downgraded, or differently scoped signed config is rejected.
+- Bootstrap tokens are single-use and absent from disk after exchange.
+- Rotation reconnects without stopping web, inventory, metrics, or cluster.
+- Credential snapshots contain only the worker account and allowed methods.
+- Public execution cannot reach loopback, link-local, RFC1918, metadata, Unix
+  sockets, or a private redirect target.
+- Revocation prevents reconnect and clears in-memory credentials.
diff --git a/docs/source-plan-migration.md b/docs/source-plan-migration.md
new file mode 100644
index 0000000..718d960
--- /dev/null
+++ b/docs/source-plan-migration.md
@@ -0,0 +1,125 @@
+# Source Plan Migration Ledger
+
+## Purpose
+
+This ledger records how worker-related documents from the RSMon control-plane
+repository were converted. The source plans are historical input, not runtime
+authority. Their worker-owned requirements now live in this repository's
+implementation documents; control-plane-owned requirements stay in RSMon.
+
+Conversion date: 2026-07-13.
+
+## Direct Worker Documents
+
+| Former RSMon document | Worker implementation document | Resolution |
+| --- | --- | --- |
+| `docs/new-worker.md` | root `README.md`, `architecture.md`, `private-workers.md` | Replaced monorepo commands with standalone binary/container/systemd operation |
+| `docs/worker-container.md` | root `README.md`, `implementation-roadmap.md` | Distribution ownership moved here; production image tags retained |
+| `docs/worker-http-settings.md` | `web-console-and-observability.md`, root `README.md` | Current port/auth behavior retained; Compose-only `WORKER_BIND_IP` clarified |
+| `docs/worker-protocol.md` | `control-plane-protocol.md` | Current tagged task envelope is normative; stale jobs/results envelopes rejected |
+| `docs/worker-cluster-quickstart.md` | `critical-check-cluster.md` | Existing cluster scaffold separated from unimplemented critical engine |
+| `docs/rsmon-distributed-api-plan.md` | `architecture.md`, `control-plane-protocol.md`, `implementation-roadmap.md` | PostgreSQL normal path retained; Redis/Rabbit/NATS queue proposal superseded |
+| `docs/plans/worker-notifier-mvp.md` | `tasks-and-notifications.md` | Final `task_envelope` and credential implementation supersede older sibling-field drafts |
+| `docs/sessions/monitor-transfer-and-admin-workers.md` | `architecture.md`, `private-workers.md` | Historical operator/session decisions reduced to current ownership and account-isolation rules |
+
+## Distributed Plans
+
+| Former RSMon document | Worker implementation document | Resolution |
+| --- | --- | --- |
+| `docs/distributed/notifications-from-worker.md` | `tasks-and-notifications.md`, `private-workers.md` | Implemented methods/statuses recorded; old credential assumptions replaced |
+| `docs/distributed/private-workers.md` | `private-workers.md` | Scheduler isolation marked current; bootstrap, signed config, and mTLS remain explicit work |
+| `docs/distributed/network-diagnostics.md` | `network-diagnostics.md` | Current confirmation separated from future rich diagnostic tasks |
+| `docs/distributed/worker-to-worker-raft.md` | `critical-check-cluster.md` | Converted to package-specific ordered implementation and release gates |
+| `docs/distributed/critical-check-amendments.md` | `critical-check-cluster.md` | No-leases, four-quorum, external-witness, and commit-before-notify rules made normative |
+| `docs/distributed/worker-web-app.md` | `web-console-and-observability.md`, `inventory.md` | Existing console recognized; future OAuth/Compose/hardware scope split into staged work |
+
+## Inventory And Host Documents
+
+| Former RSMon document | Worker implementation document | Resolution |
+| --- | --- | --- |
+| `docs/inventory-management.md` | `inventory.md` | Control-plane projection retained as dependency; worker owns collection only |
+| `docs/plans/inventory-management.md` | `inventory.md` | deploymentd and worker reports defined as source-aware complementary producers |
+| `docs/parity/rstuff-inventory.md` | `inventory.md` | rstuff enums/stream remain control-plane contracts, not worker protocol |
+| `docs/parity/deploymentd.md` | `inventory.md` | deploymentd parity becomes fixture normalization and comparison gate |
+| `docs/deploymentd-replay-rollout.md` | `inventory.md` | HTTP nonce rollout remains deploymentd-specific; worker uses authenticated typed reports |
+| `docs/integrations/rstuff-inventory-stream.md` | `inventory.md` | Valkey stream remains between control plane and rstuff, never worker transport |
+| `docs/plans/servers-and-hardware-metrics.md` | `web-console-and-observability.md`, `inventory.md` | Implemented server metric report retained; local hardware extensions staged |
+| `docs/server-observability.md` | `web-console-and-observability.md`, `control-plane-protocol.md` | Current wire fields, ownership validation, and TSDB boundary retained |
+| `docs/parity/capture.md` | `web-console-and-observability.md`, `implementation-roadmap.md` | Worker-host coverage gaps converted into staged collector work |
+
+## Diagnostics, Policy, And Product Documents
+
+| Former RSMon document | Worker implementation document | Resolution |
+| --- | --- | --- |
+| `docs/network-diagnostics.md` | `network-diagnostics.md` | Implemented confirmation/quarantine baseline retained |
+| `docs/plans/network-diagnostics-partial.md` | `network-diagnostics.md` | Partial implementation folded into explicit rich-task phases |
+| `docs/notification-task-delivery.md` | `tasks-and-notifications.md`, `control-plane-protocol.md` | Durable queue remains control-plane-owned; attempt execution remains worker-owned |
+| `docs/notification-credentials.md` | `tasks-and-notifications.md`, `private-workers.md` | SMTP/Telegram current scope retained; webhook/Mattermost gaps explicit |
+| `docs/billing-and-entitlements.md` | `private-workers.md`, `implementation-roadmap.md` | Entitlement remains a control-plane gate, never trusted from worker config |
+| `docs/plans/public-checks.md` | `private-workers.md`, `implementation-roadmap.md` | Public tasks use a separate safe grant, not widened account access |
+| `docs/plans/plans-and-billing.md` | `private-workers.md` | Plan eligibility remains control-plane policy |
+| `docs/check-kinds.md` | `tasks-and-notifications.md` | Worker-supported executor list tied to `internal/checkexec` |
+| `docs/plans/uptimerobot-parity.md` | `implementation-roadmap.md` | Worker-owned parity dependencies ordered by security prerequisites |
+| `docs/plans/krasichka.md` | none | External Krasichka runtime is not a feature of this worker binary; only protocol compatibility applies |
+
+## Operations And CI Documents
+
+| Former RSMon document | New authority | Resolution |
+| --- | --- | --- |
+| `docs/local-processes.md` | root `README.md` | Standalone development commands replace monorepo process instructions |
+| `docs/ci.md` | `.github/workflows/*.yml`, `implementation-roadmap.md` | Worker image build is owned by Gitea Actions here |
+| `docs/CI_QUICKSTART.md` | root `README.md` | Standalone make, Docker, and Compose commands are authoritative |
+
+## Resolved Conflicts
+
+### Protocol
+
+Old plans used `type: jobs/results`, separate notification siblings, and
+version-string feature detection. Current code uses `kind`, explicit branches,
+`task_envelope`, exact lease tokens, and explicit capabilities. The current
+model is authoritative.
+
+### Queue Ownership
+
+Normal task ownership is PostgreSQL with `FOR UPDATE SKIP LOCKED`. Redis
+Streams, RabbitMQ, and NATS are not worker dependencies. The critical path is
+dispatchless Raft and cannot reuse normal leases.
+
+### Credentials
+
+SMTP and Telegram use scoped pushed credential lists. Webhook and Mattermost
+are not yet equivalently persisted per account. SMS and voice are unsupported.
+No document may claim universal per-account provider isolation before those
+gaps close.
+
+### Web Console
+
+The console exists on port 27401 with Basic auth or local bcrypt. The former
+draft's port 7401, OAuth-first behavior, and entirely unimplemented status are
+obsolete. Collector lifecycle remains a real implementation gap.
+
+### Inventory
+
+deploymentd HTTP ingest is currently implemented on the control plane. Worker
+inventory is local/partial and has no upstream report yet. Future worker and
+deploymentd reports are complementary source-scoped inputs; neither may erase
+the other's observations.
+
+### Private Workers And Public Work
+
+Private workers are account-restricted. Public cross-account checks require a
+separate signed and SSRF-safe grant. Empty account scope continues to mean a
+platform-operated worker and must never be customer-selectable.
+
+### Raft Maturity
+
+Membership, persistence, snapshots, transport, and placeholder FSM entries
+exist. Production `distributed_critical` behavior does not. The control plane
+is an external witness, never a voter; provider calls remain outside Raft.
+
+## Maintenance
+
+When a new RSMon plan adds worker behavior, update this ledger and one concrete
+implementation document in the same change. Do not copy a planning draft into
+this repository without reconciling it against current packages, wire types,
+security boundaries, and tests.
diff --git a/docs/tasks-and-notifications.md b/docs/tasks-and-notifications.md
new file mode 100644
index 0000000..d935eae
--- /dev/null
+++ b/docs/tasks-and-notifications.md
@@ -0,0 +1,131 @@
+# Normal Tasks And Notifications
+
+## Normal Check Execution
+
+Normal work is centrally scheduled. The control plane owns PostgreSQL task
+rows, selection, leases, retries, deadlines, and dead-letter state. The worker
+owns only one execution attempt and its terminal report.
+
+The implemented path is:
+
+1. `internal/distworker/client.go` opens the authenticated WebSocket.
+2. `internal/distworker/runner.go` selects a recognized populated task branch.
+3. A bounded worker pool executes the task under its deadline.
+4. `internal/checkexec/exec.go` dispatches to a database-free check package.
+5. The runner queues one result and the result writer returns it to the control
+   plane with the exact lease token.
+
+Implemented check kinds are HTTP, SSL, SSH, FTP, DNS, WHOIS, BSSL, LLM,
+LLM-HTTP, ping, TCP, and UDP. Control-plane selection uses capabilities. The
+local dispatcher supports only its explicit switch cases, but strict malformed
+envelope and unsupported-kind result handling remains work package N1 rather
+than a complete fail-closed protocol response.
+
+## Queue And Shutdown Rules
+
+- Input and output queues are bounded. Backpressure must not create unbounded
+  goroutines or memory use.
+- Concurrency is supplied by the control plane but clamped by the local
+  maximum.
+- A task panic is recovered at the task boundary and reported as a failed
+  attempt; it must not kill the runner.
+- SIGTERM stops accepting tasks, allows bounded in-flight completion, attempts
+  final result delivery, and then exits.
+- WebSocket reconnect does not re-run an in-flight or completed task.
+- HTTP polling remains compatibility-only and must not become a second normal
+  scheduler.
+
+## Delegated Notifications
+
+`internal/distworker/notification.go` executes one pre-rendered delivery
+attempt. It does not render alert policy and does not own retry scheduling.
+
+| Method | State | Credential source |
+| --- | --- | --- |
+| Email | Implemented | scoped SMTP list from init/config |
+| Telegram | Implemented | scoped bot list from init/config |
+| Webhook | Implemented | current shared signing configuration |
+| Mattermost | Implemented | current shared defaults/task endpoint |
+| SMS | Explicitly unsupported | no provider selected |
+| Voice | Explicitly unsupported | no provider selected |
+
+SMS and voice return `permanent` with `unsupported_method`; they must never be
+reported as delivered or silently dropped.
+
+Worker result statuses are `delivered`, `retryable`, `permanent`, and
+`partial`. Transport failures, provider rate limits, and provider 5xx responses
+are retryable. Invalid recipients, authentication failures, malformed payloads,
+and unsupported methods are permanent. Positive provider retry-after values
+are included in the result; the control plane owns the actual retry time.
+
+## Credentials
+
+Credentials arrive in `WorkerInit.Credentials` and remain in memory. SMTP and
+Telegram tasks may request a credential ID; the worker resolves only that ID
+from the authorized pushed set. Missing IDs are permanent failures. Do not
+fall back to another account's or platform credential.
+
+The following fields must be redacted from logs and UI buffers: password,
+token, API key, signing secret, hook URL, authorization headers, and fields
+ending in `_secret`. Provider response text is bounded and must not include
+response bodies that could contain secrets.
+
+Webhook and Mattermost need persisted per-account credentials in the control
+plane before they can claim the same isolation guarantees as SMTP and
+Telegram. The worker wire shape already separates method configuration; do not
+introduce environment-global fallback for private workers.
+
+## Selfcheck Notifications
+
+System selfcheck alerts are not normal notification tasks. They allow a worker
+to notify configured system contacts when the control plane itself is
+unreachable. They use system contacts and credentials from config, local
+jitter, and local deduplication. They must not impersonate customer delivery or
+mutate control-plane message state.
+
+The local notifications page currently records selfcheck deliveries. Normal
+delegated delivery attempts should be added to the same bounded view after
+redaction, with job ID, method, status, duration, and time only.
+
+## Implementation Work Packages
+
+### N1: Runner Integration Coverage
+
+Add strict envelope/inner job-ID validation and structured unsupported-kind
+results. Add WebSocket integration tests for reconnect, task panic recovery,
+queue backpressure, result resend, stale lease behavior, malformed envelopes,
+unsupported kinds, and graceful drain. Target
+`internal/distworker/runner_protocol_test.go` and a local test WebSocket server.
+
+### N2: Complete Local Delivery Audit
+
+Record delegated notification outcomes in the bounded local notification ring.
+Never store recipient values, rendered body, credentials, or full provider
+responses in SQLite. Verify `/notifications` shows both selfcheck and delegated
+attempts without leaking secrets.
+
+### N3: Per-Account Webhook And Mattermost Credentials
+
+Once the control plane sends credential IDs and account-scoped records, require
+an exact credential match just like SMTP and Telegram. Remove global fallback
+for private workers. Add cross-account negative tests.
+
+### N4: SMS And Voice Providers
+
+Select a provider and define a credential wire type before adding an executor.
+Implementation must include timeout, retry classification, provider response
+redaction, idempotency support, and contract tests. Until all are present,
+retain the explicit permanent unsupported result.
+
+## Acceptance Tests
+
+- Every supported check kind can execute from a serialized task without a
+  database connection.
+- A stale or missing lease token cannot produce an accepted normal result.
+- One task causes at most one local execution during reconnect and duplicate
+  frame delivery.
+- Private workers cannot resolve a credential outside their authorized set.
+- SMTP 4xx/rate-limit conditions are retryable; invalid authentication is
+  permanent.
+- SMS and voice remain explicit failures until provider tests exist.
+- Logs and local UI contain no pushed credential values.
diff --git a/docs/web-console-and-observability.md b/docs/web-console-and-observability.md
new file mode 100644
index 0000000..7cbb110
--- /dev/null
+++ b/docs/web-console-and-observability.md
@@ -0,0 +1,128 @@
+# Web Console And Host Observability
+
+## Current Implementation
+
+`internal/webapp` is a functional embedded HTTP application backed by SQLite.
+It provides authenticated overview, apps, checks, notifications, logs, status,
+settings, updates, password rotation, token rotation, audit retention, and
+cluster status endpoints. Static assets and templates are embedded in the
+worker binary.
+
+The listener defaults to `WORKER_HOST=0.0.0.0` and `WORKER_PORT=27401`.
+Compose publishes it on host loopback by default through `WORKER_BIND_IP`; that
+variable is a Compose interpolation setting, not a variable read by the binary.
+
+## Authentication
+
+Two modes are implemented:
+
+| Configuration | Behavior |
+| --- | --- |
+| Both `WORKER_LOGIN` and `WORKER_PASSWORD` set | Basic-auth-backed operator sessions and `/web/api/*` access |
+| Both empty | Local bcrypt user; a one-time password is printed on first start |
+| Only one set | Startup error |
+
+Sessions are HTTP-only and SameSite strict, with a 30-minute idle and 8-hour
+absolute lifetime. State-changing browser requests require CSRF validation.
+Authenticated responses use no-store and restrictive security headers.
+
+OAuth/device authorization is not implemented. Public exposure requires a TLS
+reverse proxy and network restrictions. Do not claim that `WORKER_URL` enables
+TLS; it only advertises the externally reachable URL.
+
+## Known Runtime Gap
+
+`webapp.New` constructs `Inventory` and `Metrics`, but `Server.Start` currently
+does not start either collector and `Server.Close` does not stop them. As a
+result, production `/apps` and `/status` pages can remain empty even though
+collector unit tests pass.
+
+This is the first required web-console change:
+
+1. Start both collectors with the server context before accepting requests.
+2. Stop and wait for them during `Close` and context cancellation.
+3. Make collector start idempotent and collector failure observable.
+4. Add a server integration test that observes populated inventory and metrics
+   after startup.
+
+## Inventory View
+
+The current local collector reads Linux `/proc`, process command lines and
+working directories, TCP listeners, and resource data, then stores snapshots
+in SQLite. It does not yet provide complete grouping, Docker, Compose, systemd,
+nginx, UDP, or deploymentd-compatible inventory. The implementation plan is in
+[inventory.md](inventory.md).
+
+## Host Metrics
+
+There are two related collectors:
+
+- `internal/distworker/server_metrics.go` sends the linked server's CPU,
+  memory, disk, load, uptime, process, and network snapshot to the control
+  plane every five seconds. This path is implemented.
+- `internal/webapp/metrics.go` drives local status pages and a local ring. It
+  exists but is affected by the lifecycle gap above.
+
+The control-plane report uses interval rates for CPU and aggregate networking.
+The local metrics implementation currently computes CPU from lifetime totals;
+align it with interval deltas so local and remote views agree.
+
+## Compose Management Target
+
+Compose management remains planned and is opt-in. Implement it after read-only
+inventory is complete, under a separate `internal/composeops` package.
+
+Required constraints:
+
+- fixed stacks root, default `/opt/stacks`, with canonical path containment;
+- explicit `docker compose -f <validated-file>` invocation;
+- no shell interpolation and no free-form host command endpoint;
+- per-stack single-flight action lock;
+- async `202 Accepted` operations with bounded progress streaming;
+- viewer/operator/admin permissions and audit records;
+- typed confirmation for down, delete, and restore;
+- compose validation before save and backup before deploy;
+- Docker socket treated as root-equivalent and disabled by default.
+
+The first Compose release includes read/list/status/logs only. Mutation, file
+editing, environment editing, terminal access, backup, and restore are separate
+release gates rather than one large feature switch.
+
+## Host Status Extensions
+
+Implement in this order:
+
+1. Correct `/proc`, `statfs`, and `/proc/net/dev` interval collection.
+2. Add bounded local 24-hour one-minute aggregates.
+3. Add optional `lsblk --json` device topology.
+4. Add optional cached SMART and sensors data.
+5. Add Docker daemon and per-container status only when the socket capability
+   is enabled.
+
+Missing commands or permissions produce unavailable sections, not worker
+startup failure. All subprocesses use fixed argument arrays, deadlines, output
+limits, and allowlisted executable names.
+
+## Health And Updates
+
+`GET /healthz` is local process liveness. The image healthcheck should call
+this endpoint, not `rsmon-worker health`, because the latter probes the control
+plane `/up` endpoint. Control-plane outage is a selfcheck/readiness event, not
+proof that the worker process is dead.
+
+In-place binary replacement from the web UI is not implemented and should not
+run with ambient root access. The systemd update flow should download a signed
+artifact to a staging path and require an external privileged installer or
+explicit operator command to activate it.
+
+## Acceptance Tests
+
+- Starting `webapp.Server` produces nonempty metrics and process inventory on
+  Linux and stops every collector cleanly.
+- Both auth modes work, XOR credentials fail startup, CSRF protects writes,
+  and unauthenticated API calls fail.
+- A public bind is documented as insecure without TLS; no credential is logged.
+- `/healthz` remains healthy during a simulated control-plane outage.
+- Every host command has timeout/output-limit tests and rejects user-supplied
+  executable or path traversal.
+- Compose mutation cannot operate outside the configured stacks root.