feat(docs): Add comprehensive documentation for Vexer, Vulnerability Explorer, and Zastava modules

- Introduced AGENTS.md, README.md, TASKS.md, and implementation_plan.md for Vexer, detailing mission, responsibilities, key components, and operational notes. - Established similar documentation structure for Vulnerability Explorer and Zastava modules, including their respective workflows, integrations, and observability notes. - Created risk scoring profiles documentation outlining the core workflow, factor model, governance, and deliverables. - Ensured all modules adhere to the Aggregation-Only Contract and maintain determinism and provenance in outputs.
2025-10-30 00:09:39 +02:00
parent 3154c67978
commit 7b5bdcf4d3
503 changed files with 16136 additions and 54638 deletions
--- a/docs/modules/cli/guides/20_REFERENCE.md
+++ b/docs/modules/cli/guides/20_REFERENCE.md
@@ -0,0 +1,8 @@
+# CLI Reference (`stella --help`)
+
+> **Auto‑generated file — do not edit manually.**  
+> On every tagged release the CI pipeline runs  
+> `stella --help --markdown > docs/modules/cli/guides/20_REFERENCE.md`  
+> ensuring this document always matches the shipped binary.
+
+*(The reference will appear after the first public α release.)*
--- a/docs/modules/cli/guides/cli-reference.md
+++ b/docs/modules/cli/guides/cli-reference.md
@@ -0,0 +1,316 @@
+# CLI AOC Commands Reference
+
+> **Audience:** DevEx engineers, operators, and CI authors integrating the `stella` CLI with Aggregation-Only Contract (AOC) workflows.  
+> **Scope:** Command synopsis, options, exit codes, and offline considerations for `stella sources ingest --dry-run` and `stella aoc verify` as introduced in Sprint 19.
+
+Both commands are designed to enforce the AOC guardrails documented in the [aggregation-only reference](../../ingestion/aggregation-only-contract.md) and the [architecture overview](../../platform/architecture-overview.md). They consume Authority-issued tokens with tenant scopes and never mutate ingestion stores.
+
+---
+
+## 1 · Prerequisites
+
+- CLI version: `stella` ≥ 0.19.0 (AOC feature gate enabled).
+- Required scopes (DPoP-bound):
+  - `advisory:read` for Concelier sources.
+  - `vex:read` for Excititor sources (optional but required for VEX checks).
+  - `aoc:verify` to invoke guard verification endpoints.
+  - `tenant:select` if your deployment uses tenant switching.
+- Connectivity: direct access to Concelier/Excititor APIs or Offline Kit snapshot (see § 4).
+- Environment: set `STELLA_AUTHORITY_URL`, `STELLA_TENANT`, and export a valid OpTok via `stella auth login` or existing token cache.
+
+---
+
+## 2 · `stella sources ingest --dry-run`
+
+### 2.1 Synopsis
+
+```bash
+stella sources ingest --dry-run \
+  --source <source-key> \
+  --input <path-or-uri> \
+  [--tenant <tenant-id>] \
+  [--format json|table] \
+  [--no-color] \
+  [--output <file>]
+```
+
+### 2.2 Description
+
+Previews an ingestion write without touching MongoDB. The command loads an upstream advisory or VEX document, computes the would-write payload, runs it through the `AOCWriteGuard`, and reports any forbidden fields, provenance gaps, or idempotency issues. Use it during connector development, CI validation, or while triaging incidents.
+
+### 2.3 Options
+
+| Option | Description |
+|--------|-------------|
+| `--source <source-key>` | Logical source name (`redhat`, `ubuntu`, `osv`, etc.). Mirrors connector configuration. |
+| `--input <path-or-uri>` | Path to local CSAF/OSV/VEX file or HTTPS URI. CLI normalises transport (gzip/base64) before guard evaluation. |
+| `--tenant <tenant-id>` | Overrides default tenant for multi-tenant deployments. Mandatory when `STELLA_TENANT` is not set. |
+| `--format json|table` | Output format. `table` (default) prints summary with highlighted violations; `json` emits machine-readable report (see below). |
+| `--no-color` | Disables ANSI colour output for CI logs. |
+| `--output <file>` | Writes the JSON report to file while still printing human-readable summary to stdout. |
+
+### 2.4 Output schema (JSON)
+
+```json
+{
+  "source": "redhat",
+  "tenant": "default",
+  "guardVersion": "1.0.0",
+  "status": "ok",
+  "document": {
+    "contentHash": "sha256:…",
+    "supersedes": null,
+    "provenance": {
+      "signature": { "format": "pgp", "present": true }
+    }
+  },
+  "violations": []
+}
+```
+
+When violations exist, `status` becomes `error` and `violations` contains entries with `code` (`ERR_AOC_00x`), a short `message`, and JSON Pointer `path` values indicating offending fields.
+
+### 2.5 Exit codes
+
+| Exit code | Meaning |
+|-----------|---------|
+| `0` | Guard passed; would-write payload is AOC compliant. |
+| `11` | `ERR_AOC_001` – Forbidden field (`severity`, `cvss`, etc.) detected. |
+| `12` | `ERR_AOC_002` – Merge attempt (multiple upstream sources fused). |
+| `13` | `ERR_AOC_003` – Idempotency violation (duplicate without supersedes). |
+| `14` | `ERR_AOC_004` – Missing provenance fields. |
+| `15` | `ERR_AOC_005` – Signature/checksum mismatch. |
+| `16` | `ERR_AOC_006` – Effective findings present (Policy-only data). |
+| `17` | `ERR_AOC_007` – Unknown top-level fields / schema violation. |
+| `70` | Transport error (network, auth, malformed input). |
+
+> Exit codes map directly to the `ERR_AOC_00x` table for scripting consistency. Multiple violations yield the highest-priority code (e.g., 11 takes precedence over 14).
+
+### 2.6 Examples
+
+Dry-run a local CSAF file:
+
+```bash
+stella sources ingest --dry-run \
+  --source redhat \
+  --input ./fixtures/redhat/RHSA-2025-1234.json
+```
+
+Stream from HTTPS and emit JSON for CI:
+
+```bash
+stella sources ingest --dry-run \
+  --source osv \
+  --input https://osv.dev/vulnerability/GHSA-aaaa-bbbb \
+  --format json \
+  --output artifacts/osv-dry-run.json
+
+cat artifacts/osv-dry-run.json | jq '.violations'
+```
+
+### 2.7 Offline notes
+
+When operating in sealed/offline mode:
+
+- Use `--input` paths pointing to Offline Kit snapshots (`offline-kit/advisories/*.json`).
+- Provide `--tenant` explicitly if the offline bundle contains multiple tenants.
+- The command does not attempt network access when given a file path.
+- Store reports with `--output` to include in transfer packages for policy review.
+
+---
+
+## 3 · `stella aoc verify`
+
+### 3.1 Synopsis
+
+```bash
+stella aoc verify \
+  [--since <iso8601|duration>] \
+  [--limit <count>] \
+  [--sources <list>] \
+  [--codes <ERR_AOC_00x,...>] \
+  [--format table|json] \
+  [--export <file>] \
+  [--tenant <tenant-id>] \
+  [--no-color]
+```
+
+### 3.2 Description
+
+Replays the AOC guard against stored raw documents. By default it checks all advisories and VEX statements ingested in the last 24 hours for the active tenant, reporting totals, top violation codes, and sample documents. Use it in CI pipelines, scheduled verifications, or during incident response.
+
+### 3.3 Options
+
+| Option | Description |
+|--------|-------------|
+| `--since <value>` | Verification window. Accepts ISO 8601 timestamp (`2025-10-25T12:00:00Z`) or duration (`48h`, `7d`). Defaults to `24h`. |
+| `--limit <count>` | Maximum number of violations to display (per code). `0` means show all. Defaults to `20`. |
+| `--sources <list>` | Comma-separated list of sources (`redhat,ubuntu,osv`). Filters both advisories and VEX entries. |
+| `--codes <list>` | Restricts output to specific `ERR_AOC_00x` codes. Useful for regression tracking. |
+| `--format table|json` | `table` (default) prints summary plus top violations; `json` outputs machine-readable report identical to the `/aoc/verify` API. |
+| `--export <file>` | Writes the JSON report to disk (useful for audits/offline uploads). |
+| `--tenant <tenant-id>` | Overrides tenant context. Required for cross-tenant verifications when run by platform operators. |
+| `--no-color` | Disables ANSI colours. |
+
+`table` mode prints a summary showing the active tenant, evaluated window, counts of checked advisories/VEX statements, the active limit, total writes/violations, and whether the page was truncated. Status is colour-coded as `ok`, `violations`, or `truncated`. When violations exist the detail table lists the code, total occurrences, first sample document (`source` + `documentId` + `contentHash`), and JSON pointer path.
+
+### 3.4 Report structure (JSON)
+
+```json
+{
+  "tenant": "default",
+  "window": {
+    "from": "2025-10-25T12:00:00Z",
+    "to": "2025-10-26T12:00:00Z"
+  },
+  "checked": {
+    "advisories": 482,
+    "vex": 75
+  },
+  "violations": [
+    {
+      "code": "ERR_AOC_001",
+      "count": 2,
+      "examples": [
+        {
+          "source": "redhat",
+          "documentId": "advisory_raw:redhat:RHSA-2025:1",
+          "contentHash": "sha256:…",
+          "path": "/content/raw/cvss"
+        }
+      ]
+    }
+  ],
+  "metrics": {
+    "ingestion_write_total": 557,
+    "aoc_violation_total": 2
+  },
+  "truncated": false
+}
+```
+
+### 3.5 Exit codes
+
+| Exit code | Meaning |
+|-----------|---------|
+| `0` | Verification succeeded with zero violations. |
+| `11…17` | Same mapping as § 2.5 when violations are detected. Highest-priority code returned. |
+| `18` | Verification ran but results truncated (limit reached) – treat as warning; rerun with higher `--limit`. |
+| `70` | Transport/authentication error. |
+| `71` | CLI misconfiguration (missing tenant, invalid `--since`, etc.). |
+
+### 3.6 Examples
+
+Daily verification across all sources:
+
+```bash
+stella aoc verify --since 24h --format table
+```
+
+CI pipeline focusing on errant sources and exporting evidence:
+
+```bash
+stella aoc verify \
+  --sources redhat,ubuntu \
+  --codes ERR_AOC_001,ERR_AOC_004 \
+  --format json \
+  --limit 100 \
+  --export artifacts/aoc-verify.json
+
+jq '.violations[] | {code, count}' artifacts/aoc-verify.json
+```
+
+Air-gapped verification using Offline Kit snapshot (example script):
+
+```bash
+stella aoc verify \
+  --since 7d \
+  --format json \
+  --export /mnt/offline/aoc-verify-$(date +%F).json
+
+sha256sum /mnt/offline/aoc-verify-*.json > /mnt/offline/checksums.txt
+```
+
+### 3.7 Automation tips
+
+- Schedule with `cron` or platform scheduler and fail the job when exit code ≥ 11.
+- Pair with `stella sources ingest --dry-run` for pre-flight validation before re-enabling a paused source.
+- Push JSON exports to observability pipelines for historical tracking of violation counts.
+
+### 3.8 Offline notes
+
+- Works against Offline Kit Mongo snapshots when CLI is pointed at the local API gateway included in the bundle.
+- When fully disconnected, run against exported `aoc verify` reports generated on production and replay them using `--format json --export` (automation recipe above).
+- Include verification output in compliance packages alongside Offline Kit manifests.
+
+---
+
+## 4 · Global exit-code reference
+
+| Code | Summary |
+|------|---------|
+| `0` | Success / no violations. |
+| `11` | `ERR_AOC_001` – Forbidden field present. |
+| `12` | `ERR_AOC_002` – Merge attempt detected. |
+| `13` | `ERR_AOC_003` – Idempotency violation. |
+| `14` | `ERR_AOC_004` – Missing provenance/signature metadata. |
+| `15` | `ERR_AOC_005` – Signature/checksum mismatch. |
+| `16` | `ERR_AOC_006` – Effective findings in ingestion payload. |
+| `17` | `ERR_AOC_007` – Schema violation / unknown fields. |
+| `18` | Partial verification (limit reached). |
+| `70` | Transport or HTTP failure. |
+| `71` | CLI usage error (invalid arguments, missing tenant). |
+
+Use these codes in CI to map outcomes to build statuses or alert severities.
+
+---
+
+## 4 · `stella vuln observations` (Overlay paging)
+
+`stella vuln observations` lists raw advisory observations for downstream overlays (Graph Explorer, Policy simulations, Console). Large tenants can now page through results deterministically.
+
+| Option | Description |
+|--------|-------------|
+| `--limit <count>` | Caps the number of observations returned in a single call. Defaults to `200`; values above `500` are clamped server-side. |
+| `--cursor <token>` | Opaque continuation token produced by the previous page (`nextCursor` in JSON output). Pass it back to resume iteration. |
+
+Additional notes:
+
+- Table mode prints a hint when `hasMore` is `true`:  
+  `[yellow]More observations available. Continue with --cursor <token>[/]`.
+- JSON mode returns `nextCursor` and `hasMore` alongside the observation list so automation can loop until `hasMore` is `false`.
+- Supplying a non-positive limit falls back to the default (`200`). Invalid/expired cursors yield `400 Bad Request`; restart without `--cursor` to begin a fresh iteration.
+
+---
+
+## 5 · Related references
+
+- [Aggregation-Only Contract reference](../../ingestion/aggregation-only-contract.md)
+- [Architecture overview](../../platform/architecture-overview.md)
+- [Console AOC dashboard](../../ui/console.md)
+- [Authority scopes](../../authority/architecture.md)
+
+---
+
+## 6 · Compliance checklist
+
+- [ ] Usage documented for both table and JSON formats.
+- [ ] Exit-code mapping matches `ERR_AOC_00x` definitions and automation guidance.
+- [ ] Offline/air-gap workflow captured for both commands.
+- [ ] References to AOC architecture and console docs included.
+- [ ] Examples validated against current CLI syntax (update post-implementation).
+- [ ] Docs guild screenshot/narrative placeholder logged for release notes (pending CLI team capture).
+
+---
+
+*Last updated: 2025-10-29 (Sprint 24).*
+
+## 13. Authority configuration quick reference
+
+| Setting | Purpose | How to set |
+|---------|---------|------------|
+| `StellaOps:Authority:OperatorReason` | Incident/change description recorded with `orch:operate` tokens. | CLI flag `--Authority:OperatorReason=...` or env `STELLAOPS_ORCH_REASON`. |
+| `StellaOps:Authority:OperatorTicket` | Change/incident ticket reference paired with orchestrator control actions. | CLI flag `--Authority:OperatorTicket=...` or env `STELLAOPS_ORCH_TICKET`. |
+
+> Tokens requesting `orch:operate` will fail with `invalid_request` unless both values are present. Choose concise strings (≤256 chars for reason, ≤128 chars for ticket) and avoid sensitive data.
+
--- a/docs/modules/cli/guides/policy.md
+++ b/docs/modules/cli/guides/policy.md
@@ -0,0 +1,318 @@
+# Stella CLI — Policy Commands
+
+> **Audience:** Policy authors, reviewers, operators, and CI engineers using the `stella` CLI to interact with Policy Engine.  
+> **Supported from:** `stella` CLI ≥ 0.20.0 (Policy Engine v2 sprint line).  
+> **Prerequisites:** Authority-issued bearer token with the scopes noted per command (export `STELLA_TOKEN` or pass `--token`).
+> **2025-10-27 scope update:** CLI/CI tokens issued prior to Sprint 23 (AUTH-POLICY-23-001) must drop `policy:write`/`policy:submit`/`policy:edit` and instead request `policy:read`, `policy:author`, `policy:review`, and `policy:simulate` (plus `policy:approve`/`policy:operate`/`policy:activate` for promotion pipelines).
+
+---
+
+## 1 · Global Options & Output Modes
+
+All `stella policy *` commands honour the common CLI options:
+
+| Flag | Default | Description |
+|------|---------|-------------|
+| `--server <url>` | `https://stella.local` | Policy Engine gateway root. |
+| `--tenant <id>` | token default | Override tenant for multi-tenant installs. |
+| `--format <table\|json\|yaml>` | `table` for TTY, `json` otherwise | Output format for listings/diffs. |
+| `--output <file>` | stdout | Write full JSON payload to file. |
+| `--sealed` | false | Force sealed-mode behaviour (no outbound fetch). |
+| `--trace` | false | Emit verbose timing/log correlation info. |
+
+> **Tip:** Set `STELLA_PROFILE=policy` in CI to load saved defaults from `~/.stella/profiles/policy.toml`.
+
+---
+
+## 2 · Authoring & Drafting Commands
+
+### 2.1 `stella policy new`
+
+Create a draft policy from a template or scratch.
+
+```
+stella policy new --policy-id P-7 --name "Default Org Policy" \
+  --template baseline --output-path policies/P-7.stella
+```
+
+Options:
+
+| Flag | Description |
+|------|-------------|
+| `--policy-id` *(required)* | Stable identifier (e.g., `P-7`). |
+| `--name` | Friendly display name. |
+| `--template` | `baseline`, `serverless`, `blank`. |
+| `--from` | Start from existing version (`policyId@version`). |
+| `--open` | Launches `$EDITOR` after creation. |
+
+Writes DSL to local file and registers draft version (`status=draft`). Requires `policy:write`.
+
+### 2.2 `stella policy edit`
+
+Open an existing draft in the local editor.
+
+```
+stella policy edit P-7 --version 4
+```
+
+- Auto-checks out latest draft if `--version` omitted.
+- Saves to temp file, uploads on editor exit (unless `--no-upload`).
+- Use `--watch` to keep command alive and re-upload on every save.
+
+### 2.3 `stella policy lint`
+
+Static validation without submitting.
+
+```
+stella policy lint policies/P-7.stella --format json
+```
+
+Outputs diagnostics (line/column, code, message). Exit codes:
+
+| Code | Meaning |
+|------|---------|
+| `0` | No lint errors. |
+| `10` | Syntax/compile errors (`ERR_POL_001`). |
+| `11` | Unsupported syntax version. |
+
+### 2.4 `stella policy compile`
+
+Emits IR digest and rule summary.
+
+```
+stella policy compile P-7 --version 4
+```
+
+Returns JSON with `digest`, `rules.count`, action counts. Exit `0` success, `10` on compile errors.
+
+---
+
+## 3 · Lifecycle Workflow
+
+### 3.1 Submit
+
+```
+stella policy submit P-7 --version 4 \
+  --reviewer user:kay --reviewer group:sec-reviewers \
+  --note "Simulated against golden SBOM set" \
+  --attach sims/P-7-v4-vs-v3.json
+```
+
+Requires `policy:submit`. CLI validates that lint/compile run within 24 h and bundle attachments exist.
+
+### 3.2 Review
+
+```
+stella policy review P-7 --version 4 --approve \
+  --note "Looks good; ensure incident playbook updated."
+```
+
+- `--approve`, `--request-changes`, or `--comment`.
+- Provide `--blocking` to mark comment as blocking.
+- Requires `policy:review`.
+
+### 3.3 Approve
+
+```
+stella policy approve P-7 --version 4 \
+  --note "Determinism CI green; simulation diff attached." \
+  --attach sims/P-7-v4-vs-v3.json
+```
+
+Prompts for confirmation; refuses if approver == submitter. Requires `policy:approve`.
+
+### 3.4 Activate
+
+```
+stella policy activate P-7 --version 4 --run-now --priority high
+```
+
+- Optional `--scheduled-at 2025-10-27T02:00:00Z`.
+- Requires `policy:activate` and `policy:run`.
+
+**Options**
+
+- `--version <number>` (required) – target revision to promote.
+- `--note <text>` – record an activation note alongside the approval.
+- `--run-now` – enqueue an immediate full run after activation.
+- `--scheduled-at <timestamp>` – schedule activation for a specific UTC time (ISO-8601 format).
+- `--priority <label>` – optional scheduling priority hint (`low`, `standard`, `high`).
+- `--rollback` – mark the activation as a rollback of a previously active version.
+- `--incident <id>` – associate the activation with an incident identifier.
+
+**Exit codes**
+
+| Code | Meaning |
+|------|---------|
+| `0` | Activation completed (or policy already active). |
+| `75` | Activation recorded but awaiting a second approver. |
+
+### 3.5 Archive / Rollback
+
+```
+stella policy archive P-7 --version 3 --reason "Superseded by v4"
+stella policy activate P-7 --version 3 --rollback --incident INC-2025-104
+```
+
+---
+
+## 4 · Simulation & Runs
+
+### 4.1 Simulate
+
+```
+stella policy simulate P-7 \
+  --base 3 --candidate 4 \
+  --sbom sbom:S-42 --sbom sbom:S-318 \
+  --env exposure=internet --env sealed=false \
+  --format json --output sims/P-7-v4-vs-v3.json
+```
+
+Output fields (JSON):
+
+```json
+{
+  "diff": {
+    "added": 12,
+    "removed": 8,
+    "unchanged": 657,
+    "bySeverity": {
+      "Critical": {"up": 1, "down": 0},
+      "High": {"up": 3, "down": 4}
+    }
+  },
+  "explainUri": "blob://policy/P-7/simulations/2025-10-26.json"
+}
+```
+
+> Schema reminder: CLI commands surface objects defined in `src/Scheduler/__Libraries/StellaOps.Scheduler.Models/docs/SCHED-MODELS-20-001-POLICY-RUNS.md`; use the samples in `samples/api/scheduler/` for contract validation when extending output parsing.
+
+Exit codes:
+
+| Code | Meaning |
+|------|---------|
+| `0` | Simulation succeeded; diffs informational. |
+| `20` | Blocking delta (`--fail-on-diff` triggered). |
+| `21` | Simulation input missing (`ERR_POL_003`). |
+| `22` | Determinism guard (`ERR_POL_004`). |
+| `23` | API/permission error (`ERR_POL_002`, `ERR_POL_005`). |
+
+### 4.2 Run
+
+```
+stella policy run P-7 --mode full \
+  --sbom sbom:S-42 --env exposure=internal-only \
+  --wait --watch
+```
+
+Options:
+
+| Flag | Description |
+|------|-------------|
+| `--mode` | `full` or `incremental` (default incremental). |
+| `--sbom` | Explicit SBOM IDs (optional). |
+| `--priority` | `normal`, `high`, `emergency`. |
+| `--wait` | Poll run status until completion. |
+| `--watch` | Stream progress events (requires TTY). |
+
+`stella policy run status <runId>` retrieves run metadata.  
+`stella policy run list --status failed --limit 20` returns recent runs.
+
+### 4.3 Replay & Cancel
+
+```
+stella policy run replay run:P-7:2025-10-26:auto --output bundles/replay.tgz
+stella policy run cancel run:P-7:2025-10-26:auto
+```
+
+Replay downloads sealed bundle for deterministic verification.
+
+### 4.4 Schema artefacts for CLI validation
+
+- CI publishes canonical JSON Schema exports for `PolicyRunRequest`, `PolicyRunStatus`, `PolicyDiffSummary`, and `PolicyExplainTrace` as the `policy-schema-exports` artifact (see `.gitea/workflows/build-test-deploy.yml`).
+- Each run writes the files to `artifacts/policy-schemas/<commit>/` and stores a unified diff (`policy-schema-diff.patch`) comparing them with the tracked baseline in `docs/schemas/`.
+- Schema changes trigger an alert in Slack `#policy-engine` via the `POLICY_ENGINE_SCHEMA_WEBHOOK` secret so CLI maintainers know to refresh fixtures or validation rules.
+- Consume these artefacts in CLI tests to keep payload validation aligned without committing generated files into the repo.
+
+---
+
+## 5 · Findings & Explainability
+
+### 5.1 List Findings
+
+```
+stella findings ls --policy P-7 \
+  --sbom sbom:S-42 \
+  --status affected --severity High,Critical \
+  --since 2025-10-01T00:00:00Z \
+  --page 2 --page-size 100 \
+  --format table
+```
+
+Common flags:
+
+| Flag | Description |
+|------|-------------|
+| `--sbom` | Repeatable filter for SBOM identifiers. |
+| `--status` | Repeatable filter (`affected`, `quieted`, `mitigated`, `not_affected`, etc.). |
+| `--severity` | Repeatable filter using normalized labels (`Critical`, `High`, `Medium`, `Low`, `Unknown`). |
+| `--since` | Return findings updated on/after the ISO-8601 timestamp. |
+| `--cursor` | Resume listing using the opaque token from a prior page. |
+| `--page`, `--page-size` | Page-based pagination (page >=1, size <=500; falls back to backend defaults). |
+| `--output` | Persist JSON payload to disk (implied JSON rendering). |
+| `--format` | `table` (default for TTY) or `json`. |
+
+### 5.2 Fetch Explain
+
+```
+stella findings explain --policy P-7 \
+  P-7:S-42:pkg:npm/lodash@4.17.21:CVE-2021-23337 \
+  --mode verbose \
+  --format json --output explains/lodash.json
+```
+
+Outputs ordered rule hits, inputs, evidence snapshots, and sealed-mode hints. Supported `--mode` values mirror API contracts (for example `summary`, `verbose`); omit to use backend default.
+
+---
+
+## 6 · Exit Codes Summary
+
+| Exit code | Description | Typical ERR codes |
+|-----------|-------------|-------------------|
+| `0` | Success (command completed, warnings only). | — |
+| `10` | DSL syntax/compile failure. | `ERR_POL_001` |
+| `11` | Unsupported DSL version / schema mismatch. | `ERR_POL_001` |
+| `12` | Approval/rbac failure. | `ERR_POL_002`, `ERR_POL_005` |
+| `20` | Simulation diff exceeded thresholds (`--fail-on-diff`). | — |
+| `21` | Required inputs missing (SBOM/advisory/VEX). | `ERR_POL_003` |
+| `22` | Determinism guard triggered. | `ERR_POL_004` |
+| `23` | Run canceled or timed out. | `ERR_POL_006` |
+| `30` | Network/transport error (non-HTTP success). | — |
+| `64` | CLI usage error (invalid flag/argument). | — |
+
+All non-zero exits emit structured error envelope on stderr when `--format json` or `STELLA_JSON_ERRORS=1`.
+
+---
+
+## 7 · Offline & Air-Gap Usage
+
+- Use `--sealed` to ensure commands avoid outbound calls; required for sealed enclaves.
+- `stella policy bundle export --policy P-7 --version 4 --output bundles/policy-P-7-v4.bundle` pairs with Offline Kit import.
+- Replay bundles (`run replay`) are DSSE-signed; verify with `stella offline verify`.
+- Store credentials in `~/.stella/offline.toml` for non-interactive air-gapped pipelines.
+
+---
+
+## 8 · Compliance Checklist
+
+- [ ] **Help text synced:** `stella policy --help` matches documented flags/examples (update during release pipeline).
+- [ ] **Exit codes mapped:** Table above reflects CLI implementation and CI asserts mapping for `ERR_POL_*`.
+- [ ] **JSON schemas verified:** Example payloads validated against OpenAPI/SDK contracts before publishing. (_CI now exports canonical schemas as `policy-schema-exports`; wire tests to consume them._)
+- [ ] **Scope guidance present:** Each command lists required Authority scopes.
+- [ ] **Offline guidance included:** Sealed-mode steps and bundle workflows documented.
+- [ ] **Cross-links tested:** Links to DSL, lifecycle, runs, and API docs resolve locally (`yarn docs:lint`).
+- [ ] **Examples no-op safe:** Command examples either read-only or use placeholders (no destructive defaults).
+
+---
+
+*Last updated: 2025-10-27 (Sprint 20).*