coderifts 4.9.0 → 4.11.0

package/CHANGELOG.md

@@ -1,5 +1,13 @@
 # Changelog
 
+## 4.10.0
+
+### Added
+- **`coderifts init --agents --strict`** writes the STRICT contract-gate
+  workflow: `require-verified-monitoring: 'true'` (plus attestation /
+  keyring inputs). The default template is unchanged (monitoring stays
+  commented / false).
+
 ## 4.9.0
 
 ### Fixed

package/README.md

@@ -75,6 +75,7 @@ coderifts init --agents --hosts=cursor   # Cursor only
 coderifts init --agents --hosts=copilot  # GitHub Copilot / VS Code only
 coderifts init --agents --no-hook        # skip host hook install
 coderifts init --agents --no-workflow    # skip .github/workflows/coderifts.yml
+coderifts init --agents --strict         # STRICT: require-verified-monitoring: true
 coderifts init --agents --check          # which of the four pieces are present
 ```
 
@@ -83,7 +84,7 @@ coderifts init --agents --check          # which of the four pieces are present
 | **MCP config** | `.mcp.json` (Claude Code), `.cursor/mcp.json` (Cursor), `.vscode/mcp.json` (Copilot / VS Code) | Three shapes, never normalised into one: Claude `mcpServers` + `type: "http"` + `url`; Cursor `mcpServers` + `url` only (no `type`); Copilot **`servers`** (not `mcpServers`) + `type: "http"` + `url`. Never `httpUrl` (that is Gemini, extension-level — this command does not write it). |
 | **Agent rules** | generated host files via `agent-setup`'s writer | `AGENTS.md` is always written. Claude also gets `CLAUDE.md`; Cursor also gets `.cursor/rules/coderifts.mdc`; Copilot also gets `.github/copilot-instructions.md`. `--hosts=all` adds LangGraph and OpenAI instructions. (`coderifts init agent` is a different command — it writes a `.coderifts.yml` template.) |
 | **Host hook** | `.claude/settings.json`, `.cursor/hooks.json` | Reuses `hook install --claude` / `--cursor` (JSON-merge, backup, idempotent). Copilot shares Claude's `.claude/settings.json` — VS Code reads it by default. |
-| **CI workflow** | `.github/workflows/coderifts.yml` | `coderifts/contract-gate@v0`. `require-verified-monitoring` is commented (honest default: **false**) |
+| **CI workflow** | `.github/workflows/coderifts.yml` | `coderifts/contract-gate@v0`. Default: `require-verified-monitoring` commented (honest default: **false**). `--strict`: the same file with `require-verified-monitoring: 'true'` |
 
 **Next manual step:** add `CODERIFTS_API_KEY` to the repository secrets. The
 CLI never writes a key anywhere.

package/bin/coderifts.js

@@ -95,6 +95,7 @@ program
   .option('--hosts <list>', 'With --agents: claude, cursor, copilot, or all (default: all)', 'all')
   .option('--no-hook', 'With --agents: skip host hook install')
   .option('--no-workflow', 'With --agents: skip writing .github/workflows/coderifts.yml')
+  .option('--strict', 'With --agents: STRICT workflow (require-verified-monitoring: true)')
   .option('--dry-run', 'With --agents: print the plan without writing')
   .option('--check', 'With --agents: report which of the four pieces are present (no write)')
   .option('--out <dir>', 'With --agents: target directory (default: cwd)')
@@ -118,6 +119,7 @@ program
         dryRun: !!options.dryRun,
         check: !!options.check,
         out: options.out,
+        strict: !!options.strict,
       });
       if (result && typeof result.exitCode === 'number') {
         process.exitCode = result.exitCode;

package/dist/cli.js

@@ -3028,7 +3028,7 @@ var require_package = __commonJS({
   "package.json"(exports2, module2) {
     module2.exports = {
       name: "coderifts",
-      version: "4.9.0",
+      version: "4.11.0",
       description: "Detect breaking API changes from the command line. Works locally or with the CodeRifts cloud API.",
       author: "CodeRifts <hello@coderifts.com>",
       license: "MIT",
@@ -209173,16 +209173,18 @@ var require_agent_host_files_embedded = __commonJS({
       ".github/copilot-instructions.md",
       "coderifts-langgraph-policy.js",
       "openai-agent-instructions.md",
+      "skills/api-governance/SKILL.md",
       ".claude/settings.json",
       ".cursor/hooks.json"
     ]);
     var AGENT_HOST_FILES = Object.freeze({
-      "AGENTS.md": '<!-- GENERATED from CodeRifts agent-host-rule (single source). Do not edit rule sentences by hand; run: node scripts/generate-agent-host-files.js -->\n\n# CodeRifts agent rules\n\nCall `preflight_change_set` before merge, deploy, publish, or tool registration when a contract artifact changed (OpenAPI/Swagger, GraphQL, gRPC/protobuf, AsyncAPI, MCP manifest, or agent tool schemas).\n\nSend the complete base-to-head change set: every changed contract artifact with full before and after content. Do not send a single-file subset when other contract files also change.\n\nBranch on `execution_action` only. Do not branch on `decision` and do not branch on `safe_for_agent` (not_for_control_flow_use_execution_action). Canonical `execution_action` values: CONTINUE, CONTINUE_WITH_MONITORING, REQUEST_APPROVAL, STOP.\n\nAn unrecognised `execution_action` is not permission: fail closed (halt or re-preflight). Well-known code: `not_permission_fail_closed`.\n\n`CONTINUE_WITH_MONITORING` requires a wired monitoring sink (`monitoringSinkWired`). It is not "proceed with caution" without monitoring.\n\nDo not call CodeRifts tools for a documentation-only change (README, guides, comments) with no contract artifact content change.\n\nIf you already hold a chain receipt and only need authenticity/lifecycle: `verify_receipt`. If you need a past decision by id: `get_decision_details`. Neither replaces preflight for a new change set.\n\nA receipt authorizes ONE operation: a merge receipt does not authorize a deploy. Before a different operation (deploy, publish), call `preflight_change_set` with `context.operation` set to that operation \u2014 reusing a differently-scoped receipt is not permitted and will fail at the gate.\n\nA stale or superseded receipt on a changed head requires a NEW preflight \u2014 `verify_receipt` cannot re-diff.\n\nFor mutating tools, put only the guarded version in the agent\'s tool table; keep the raw handler host-only and unreachable from that table. How you name tools is yours \u2014 this is a reachability property, not a product rename of host tools. CodeRifts cannot see or stop a raw call the host makes outside the table it returns; adopt this as a host convention, not as a guarantee from the package.\n\nCodeRifts reports a governance decision and `execution_action`; it does not by itself block merges. Blocking requires separate repository configuration (required status checks, enforcement) that this rule file does not set.\n\nTo act (mutate a contract, merge, deploy, or publish): call `preflight_change_set` with `preflight_mode` authorize. Analyze is informational (`may_execute` is always false) and is not permission. Read `execution_action` on the `decision_result` envelope.\n\nBefore acting under a held receipt: call `verify_receipt` with the intended `context` (operation, environment, repository, branch, pull_request) for THIS attempt. Do not act on a receipt whose scope does not match.\n\nAct only when `currently_authorized` is true (`control_envelope.receipt_view.currently_authorized`). A valid-looking token is not permission if `currently_authorized` is false or omitted.\n\nCommit / CAS evidence is a separate measurement (`commit_observation` on GuardOutcome). It is not a substitute for authorize + `currently_authorized`. Production hosts that want the fail-closed conjunction lock it with `profile: ENFORCING_STRICT` on withCodeRifts.\n\nIf the host requests an execution grant (opt-in `include_execution_grant`), the grant is bound to operation + target + after-payload (`scope_hash`) and is short-lived \u2014 never reuse it after the after-payload changes.\n\nAn ATOMIC-profile grant carries `state_nonce` and is single-use at the executor \u2014 if the executor has consumed the nonce, re-preflight; do not retry the same grant.\n\nWith a proven tenant\u2194repo binding you may request `derivation:"server"` instead of assembling `artifacts[]` yourself (`context.repository` + `context.base` + `context.head` required; caller-supplied artifacts are rejected on that path).\n\nA commit is only proven when an executor attestation verifies (customer-held executor key, `cas_evidence: executor_attested`); otherwise say "authorized, commit not proven".\n',
-      "CLAUDE.md": '<!-- GENERATED from CodeRifts agent-host-rule (single source). Do not edit rule sentences by hand; run: node scripts/generate-agent-host-files.js -->\n\n# CodeRifts\n\nCall `preflight_change_set` before merge, deploy, publish, or tool registration when a contract artifact changed (OpenAPI/Swagger, GraphQL, gRPC/protobuf, AsyncAPI, MCP manifest, or agent tool schemas).\n\nSend the complete base-to-head change set: every changed contract artifact with full before and after content. Do not send a single-file subset when other contract files also change.\n\nBranch on `execution_action` only. Do not branch on `decision` and do not branch on `safe_for_agent` (not_for_control_flow_use_execution_action). Canonical `execution_action` values: CONTINUE, CONTINUE_WITH_MONITORING, REQUEST_APPROVAL, STOP.\n\nAn unrecognised `execution_action` is not permission: fail closed (halt or re-preflight). Well-known code: `not_permission_fail_closed`.\n\n`CONTINUE_WITH_MONITORING` requires a wired monitoring sink (`monitoringSinkWired`). It is not "proceed with caution" without monitoring.\n\nDo not call CodeRifts tools for a documentation-only change (README, guides, comments) with no contract artifact content change.\n\nIf you already hold a chain receipt and only need authenticity/lifecycle: `verify_receipt`. If you need a past decision by id: `get_decision_details`. Neither replaces preflight for a new change set.\n\nA receipt authorizes ONE operation: a merge receipt does not authorize a deploy. Before a different operation (deploy, publish), call `preflight_change_set` with `context.operation` set to that operation \u2014 reusing a differently-scoped receipt is not permitted and will fail at the gate.\n\nA stale or superseded receipt on a changed head requires a NEW preflight \u2014 `verify_receipt` cannot re-diff.\n\nFor mutating tools, put only the guarded version in the agent\'s tool table; keep the raw handler host-only and unreachable from that table. How you name tools is yours \u2014 this is a reachability property, not a product rename of host tools. CodeRifts cannot see or stop a raw call the host makes outside the table it returns; adopt this as a host convention, not as a guarantee from the package.\n\nCodeRifts reports a governance decision and `execution_action`; it does not by itself block merges. Blocking requires separate repository configuration (required status checks, enforcement) that this rule file does not set.\n\nTo act (mutate a contract, merge, deploy, or publish): call `preflight_change_set` with `preflight_mode` authorize. Analyze is informational (`may_execute` is always false) and is not permission. Read `execution_action` on the `decision_result` envelope.\n\nBefore acting under a held receipt: call `verify_receipt` with the intended `context` (operation, environment, repository, branch, pull_request) for THIS attempt. Do not act on a receipt whose scope does not match.\n\nAct only when `currently_authorized` is true (`control_envelope.receipt_view.currently_authorized`). A valid-looking token is not permission if `currently_authorized` is false or omitted.\n\nCommit / CAS evidence is a separate measurement (`commit_observation` on GuardOutcome). It is not a substitute for authorize + `currently_authorized`. Production hosts that want the fail-closed conjunction lock it with `profile: ENFORCING_STRICT` on withCodeRifts.\n\nIf the host requests an execution grant (opt-in `include_execution_grant`), the grant is bound to operation + target + after-payload (`scope_hash`) and is short-lived \u2014 never reuse it after the after-payload changes.\n\nAn ATOMIC-profile grant carries `state_nonce` and is single-use at the executor \u2014 if the executor has consumed the nonce, re-preflight; do not retry the same grant.\n\nWith a proven tenant\u2194repo binding you may request `derivation:"server"` instead of assembling `artifacts[]` yourself (`context.repository` + `context.base` + `context.head` required; caller-supplied artifacts are rejected on that path).\n\nA commit is only proven when an executor attestation verifies (customer-held executor key, `cas_evidence: executor_attested`); otherwise say "authorized, commit not proven".\n',
-      ".cursor/rules/coderifts.mdc": '---\ndescription: CodeRifts API governance \u2014 when to preflight and how to branch\nglobs:\nalwaysApply: true\n---\n\n<!-- GENERATED from CodeRifts agent-host-rule (single source). Do not edit rule sentences by hand; run: node scripts/generate-agent-host-files.js -->\n\n# CodeRifts\n\nCall `preflight_change_set` before merge, deploy, publish, or tool registration when a contract artifact changed (OpenAPI/Swagger, GraphQL, gRPC/protobuf, AsyncAPI, MCP manifest, or agent tool schemas).\n\nSend the complete base-to-head change set: every changed contract artifact with full before and after content. Do not send a single-file subset when other contract files also change.\n\nBranch on `execution_action` only. Do not branch on `decision` and do not branch on `safe_for_agent` (not_for_control_flow_use_execution_action). Canonical `execution_action` values: CONTINUE, CONTINUE_WITH_MONITORING, REQUEST_APPROVAL, STOP.\n\nAn unrecognised `execution_action` is not permission: fail closed (halt or re-preflight). Well-known code: `not_permission_fail_closed`.\n\n`CONTINUE_WITH_MONITORING` requires a wired monitoring sink (`monitoringSinkWired`). It is not "proceed with caution" without monitoring.\n\nDo not call CodeRifts tools for a documentation-only change (README, guides, comments) with no contract artifact content change.\n\nIf you already hold a chain receipt and only need authenticity/lifecycle: `verify_receipt`. If you need a past decision by id: `get_decision_details`. Neither replaces preflight for a new change set.\n\nA receipt authorizes ONE operation: a merge receipt does not authorize a deploy. Before a different operation (deploy, publish), call `preflight_change_set` with `context.operation` set to that operation \u2014 reusing a differently-scoped receipt is not permitted and will fail at the gate.\n\nA stale or superseded receipt on a changed head requires a NEW preflight \u2014 `verify_receipt` cannot re-diff.\n\nFor mutating tools, put only the guarded version in the agent\'s tool table; keep the raw handler host-only and unreachable from that table. How you name tools is yours \u2014 this is a reachability property, not a product rename of host tools. CodeRifts cannot see or stop a raw call the host makes outside the table it returns; adopt this as a host convention, not as a guarantee from the package.\n\nCodeRifts reports a governance decision and `execution_action`; it does not by itself block merges. Blocking requires separate repository configuration (required status checks, enforcement) that this rule file does not set.\n\nTo act (mutate a contract, merge, deploy, or publish): call `preflight_change_set` with `preflight_mode` authorize. Analyze is informational (`may_execute` is always false) and is not permission. Read `execution_action` on the `decision_result` envelope.\n\nBefore acting under a held receipt: call `verify_receipt` with the intended `context` (operation, environment, repository, branch, pull_request) for THIS attempt. Do not act on a receipt whose scope does not match.\n\nAct only when `currently_authorized` is true (`control_envelope.receipt_view.currently_authorized`). A valid-looking token is not permission if `currently_authorized` is false or omitted.\n\nCommit / CAS evidence is a separate measurement (`commit_observation` on GuardOutcome). It is not a substitute for authorize + `currently_authorized`. Production hosts that want the fail-closed conjunction lock it with `profile: ENFORCING_STRICT` on withCodeRifts.\n\nIf the host requests an execution grant (opt-in `include_execution_grant`), the grant is bound to operation + target + after-payload (`scope_hash`) and is short-lived \u2014 never reuse it after the after-payload changes.\n\nAn ATOMIC-profile grant carries `state_nonce` and is single-use at the executor \u2014 if the executor has consumed the nonce, re-preflight; do not retry the same grant.\n\nWith a proven tenant\u2194repo binding you may request `derivation:"server"` instead of assembling `artifacts[]` yourself (`context.repository` + `context.base` + `context.head` required; caller-supplied artifacts are rejected on that path).\n\nA commit is only proven when an executor attestation verifies (customer-held executor key, `cas_evidence: executor_attested`); otherwise say "authorized, commit not proven".\n',
-      ".github/copilot-instructions.md": '<!-- GENERATED from CodeRifts agent-host-rule (single source). Do not edit rule sentences by hand; run: node scripts/generate-agent-host-files.js -->\n\n# CodeRifts instructions for GitHub Copilot\n\nCall `preflight_change_set` before merge, deploy, publish, or tool registration when a contract artifact changed (OpenAPI/Swagger, GraphQL, gRPC/protobuf, AsyncAPI, MCP manifest, or agent tool schemas).\n\nSend the complete base-to-head change set: every changed contract artifact with full before and after content. Do not send a single-file subset when other contract files also change.\n\nBranch on `execution_action` only. Do not branch on `decision` and do not branch on `safe_for_agent` (not_for_control_flow_use_execution_action). Canonical `execution_action` values: CONTINUE, CONTINUE_WITH_MONITORING, REQUEST_APPROVAL, STOP.\n\nAn unrecognised `execution_action` is not permission: fail closed (halt or re-preflight). Well-known code: `not_permission_fail_closed`.\n\n`CONTINUE_WITH_MONITORING` requires a wired monitoring sink (`monitoringSinkWired`). It is not "proceed with caution" without monitoring.\n\nDo not call CodeRifts tools for a documentation-only change (README, guides, comments) with no contract artifact content change.\n\nIf you already hold a chain receipt and only need authenticity/lifecycle: `verify_receipt`. If you need a past decision by id: `get_decision_details`. Neither replaces preflight for a new change set.\n\nA receipt authorizes ONE operation: a merge receipt does not authorize a deploy. Before a different operation (deploy, publish), call `preflight_change_set` with `context.operation` set to that operation \u2014 reusing a differently-scoped receipt is not permitted and will fail at the gate.\n\nA stale or superseded receipt on a changed head requires a NEW preflight \u2014 `verify_receipt` cannot re-diff.\n\nFor mutating tools, put only the guarded version in the agent\'s tool table; keep the raw handler host-only and unreachable from that table. How you name tools is yours \u2014 this is a reachability property, not a product rename of host tools. CodeRifts cannot see or stop a raw call the host makes outside the table it returns; adopt this as a host convention, not as a guarantee from the package.\n\nCodeRifts reports a governance decision and `execution_action`; it does not by itself block merges. Blocking requires separate repository configuration (required status checks, enforcement) that this rule file does not set.\n\nTo act (mutate a contract, merge, deploy, or publish): call `preflight_change_set` with `preflight_mode` authorize. Analyze is informational (`may_execute` is always false) and is not permission. Read `execution_action` on the `decision_result` envelope.\n\nBefore acting under a held receipt: call `verify_receipt` with the intended `context` (operation, environment, repository, branch, pull_request) for THIS attempt. Do not act on a receipt whose scope does not match.\n\nAct only when `currently_authorized` is true (`control_envelope.receipt_view.currently_authorized`). A valid-looking token is not permission if `currently_authorized` is false or omitted.\n\nCommit / CAS evidence is a separate measurement (`commit_observation` on GuardOutcome). It is not a substitute for authorize + `currently_authorized`. Production hosts that want the fail-closed conjunction lock it with `profile: ENFORCING_STRICT` on withCodeRifts.\n\nIf the host requests an execution grant (opt-in `include_execution_grant`), the grant is bound to operation + target + after-payload (`scope_hash`) and is short-lived \u2014 never reuse it after the after-payload changes.\n\nAn ATOMIC-profile grant carries `state_nonce` and is single-use at the executor \u2014 if the executor has consumed the nonce, re-preflight; do not retry the same grant.\n\nWith a proven tenant\u2194repo binding you may request `derivation:"server"` instead of assembling `artifacts[]` yourself (`context.repository` + `context.base` + `context.head` required; caller-supplied artifacts are rejected on that path).\n\nA commit is only proven when an executor attestation verifies (customer-held executor key, `cas_evidence: executor_attested`); otherwise say "authorized, commit not proven".\n',
-      "coderifts-langgraph-policy.js": '// GENERATED from CodeRifts agent-host-rule (single source). Do not edit rule sentences by hand; run: node scripts/generate-agent-host-files.js\n// System policy string for a LangGraph (or similar) agent. Content is generated;\n// identical rule sentences to AGENTS.md / other formats.\n\'use strict\';\n\nmodule.exports = "Call `preflight_change_set` before merge, deploy, publish, or tool registration when a contract artifact changed (OpenAPI/Swagger, GraphQL, gRPC/protobuf, AsyncAPI, MCP manifest, or agent tool schemas).\\n\\nSend the complete base-to-head change set: every changed contract artifact with full before and after content. Do not send a single-file subset when other contract files also change.\\n\\nBranch on `execution_action` only. Do not branch on `decision` and do not branch on `safe_for_agent` (not_for_control_flow_use_execution_action). Canonical `execution_action` values: CONTINUE, CONTINUE_WITH_MONITORING, REQUEST_APPROVAL, STOP.\\n\\nAn unrecognised `execution_action` is not permission: fail closed (halt or re-preflight). Well-known code: `not_permission_fail_closed`.\\n\\n`CONTINUE_WITH_MONITORING` requires a wired monitoring sink (`monitoringSinkWired`). It is not \\"proceed with caution\\" without monitoring.\\n\\nDo not call CodeRifts tools for a documentation-only change (README, guides, comments) with no contract artifact content change.\\n\\nIf you already hold a chain receipt and only need authenticity/lifecycle: `verify_receipt`. If you need a past decision by id: `get_decision_details`. Neither replaces preflight for a new change set.\\n\\nA receipt authorizes ONE operation: a merge receipt does not authorize a deploy. Before a different operation (deploy, publish), call `preflight_change_set` with `context.operation` set to that operation \u2014 reusing a differently-scoped receipt is not permitted and will fail at the gate.\\n\\nA stale or superseded receipt on a changed head requires a NEW preflight \u2014 `verify_receipt` cannot re-diff.\\n\\nFor mutating tools, put only the guarded version in the agent\'s tool table; keep the raw handler host-only and unreachable from that table. How you name tools is yours \u2014 this is a reachability property, not a product rename of host tools. CodeRifts cannot see or stop a raw call the host makes outside the table it returns; adopt this as a host convention, not as a guarantee from the package.\\n\\nCodeRifts reports a governance decision and `execution_action`; it does not by itself block merges. Blocking requires separate repository configuration (required status checks, enforcement) that this rule file does not set.\\n\\nTo act (mutate a contract, merge, deploy, or publish): call `preflight_change_set` with `preflight_mode` authorize. Analyze is informational (`may_execute` is always false) and is not permission. Read `execution_action` on the `decision_result` envelope.\\n\\nBefore acting under a held receipt: call `verify_receipt` with the intended `context` (operation, environment, repository, branch, pull_request) for THIS attempt. Do not act on a receipt whose scope does not match.\\n\\nAct only when `currently_authorized` is true (`control_envelope.receipt_view.currently_authorized`). A valid-looking token is not permission if `currently_authorized` is false or omitted.\\n\\nCommit / CAS evidence is a separate measurement (`commit_observation` on GuardOutcome). It is not a substitute for authorize + `currently_authorized`. Production hosts that want the fail-closed conjunction lock it with `profile: ENFORCING_STRICT` on withCodeRifts.\\n\\nIf the host requests an execution grant (opt-in `include_execution_grant`), the grant is bound to operation + target + after-payload (`scope_hash`) and is short-lived \u2014 never reuse it after the after-payload changes.\\n\\nAn ATOMIC-profile grant carries `state_nonce` and is single-use at the executor \u2014 if the executor has consumed the nonce, re-preflight; do not retry the same grant.\\n\\nWith a proven tenant\u2194repo binding you may request `derivation:\\"server\\"` instead of assembling `artifacts[]` yourself (`context.repository` + `context.base` + `context.head` required; caller-supplied artifacts are rejected on that path).\\n\\nA commit is only proven when an executor attestation verifies (customer-held executor key, `cas_evidence: executor_attested`); otherwise say \\"authorized, commit not proven\\".";\n',
-      "openai-agent-instructions.md": '<!-- GENERATED from CodeRifts agent-host-rule (single source). Do not edit rule sentences by hand; run: node scripts/generate-agent-host-files.js -->\n\n# CodeRifts agent instructions\n\nCall `preflight_change_set` before merge, deploy, publish, or tool registration when a contract artifact changed (OpenAPI/Swagger, GraphQL, gRPC/protobuf, AsyncAPI, MCP manifest, or agent tool schemas).\n\nSend the complete base-to-head change set: every changed contract artifact with full before and after content. Do not send a single-file subset when other contract files also change.\n\nBranch on `execution_action` only. Do not branch on `decision` and do not branch on `safe_for_agent` (not_for_control_flow_use_execution_action). Canonical `execution_action` values: CONTINUE, CONTINUE_WITH_MONITORING, REQUEST_APPROVAL, STOP.\n\nAn unrecognised `execution_action` is not permission: fail closed (halt or re-preflight). Well-known code: `not_permission_fail_closed`.\n\n`CONTINUE_WITH_MONITORING` requires a wired monitoring sink (`monitoringSinkWired`). It is not "proceed with caution" without monitoring.\n\nDo not call CodeRifts tools for a documentation-only change (README, guides, comments) with no contract artifact content change.\n\nIf you already hold a chain receipt and only need authenticity/lifecycle: `verify_receipt`. If you need a past decision by id: `get_decision_details`. Neither replaces preflight for a new change set.\n\nA receipt authorizes ONE operation: a merge receipt does not authorize a deploy. Before a different operation (deploy, publish), call `preflight_change_set` with `context.operation` set to that operation \u2014 reusing a differently-scoped receipt is not permitted and will fail at the gate.\n\nA stale or superseded receipt on a changed head requires a NEW preflight \u2014 `verify_receipt` cannot re-diff.\n\nFor mutating tools, put only the guarded version in the agent\'s tool table; keep the raw handler host-only and unreachable from that table. How you name tools is yours \u2014 this is a reachability property, not a product rename of host tools. CodeRifts cannot see or stop a raw call the host makes outside the table it returns; adopt this as a host convention, not as a guarantee from the package.\n\nCodeRifts reports a governance decision and `execution_action`; it does not by itself block merges. Blocking requires separate repository configuration (required status checks, enforcement) that this rule file does not set.\n\nTo act (mutate a contract, merge, deploy, or publish): call `preflight_change_set` with `preflight_mode` authorize. Analyze is informational (`may_execute` is always false) and is not permission. Read `execution_action` on the `decision_result` envelope.\n\nBefore acting under a held receipt: call `verify_receipt` with the intended `context` (operation, environment, repository, branch, pull_request) for THIS attempt. Do not act on a receipt whose scope does not match.\n\nAct only when `currently_authorized` is true (`control_envelope.receipt_view.currently_authorized`). A valid-looking token is not permission if `currently_authorized` is false or omitted.\n\nCommit / CAS evidence is a separate measurement (`commit_observation` on GuardOutcome). It is not a substitute for authorize + `currently_authorized`. Production hosts that want the fail-closed conjunction lock it with `profile: ENFORCING_STRICT` on withCodeRifts.\n\nIf the host requests an execution grant (opt-in `include_execution_grant`), the grant is bound to operation + target + after-payload (`scope_hash`) and is short-lived \u2014 never reuse it after the after-payload changes.\n\nAn ATOMIC-profile grant carries `state_nonce` and is single-use at the executor \u2014 if the executor has consumed the nonce, re-preflight; do not retry the same grant.\n\nWith a proven tenant\u2194repo binding you may request `derivation:"server"` instead of assembling `artifacts[]` yourself (`context.repository` + `context.base` + `context.head` required; caller-supplied artifacts are rejected on that path).\n\nA commit is only proven when an executor attestation verifies (customer-held executor key, `cas_evidence: executor_attested`); otherwise say "authorized, commit not proven".\n',
+      "AGENTS.md": '<!-- GENERATED from CodeRifts agent-host-rule (single source). Do not edit rule sentences by hand; run: node scripts/generate-agent-host-files.js -->\n\n# CodeRifts agent rules\n\nCall `preflight_change_set` before merge, deploy, publish, or tool registration when a contract artifact changed (OpenAPI/Swagger, GraphQL, gRPC/protobuf, AsyncAPI, MCP manifest, or agent tool schemas).\n\nSend the complete base-to-head change set: every changed contract artifact with full before and after content. Do not send a single-file subset when other contract files also change.\n\nBranch on `execution_action` only. Do not branch on `decision` and do not branch on `safe_for_agent` (not_for_control_flow_use_execution_action). Canonical `execution_action` values: CONTINUE, CONTINUE_WITH_MONITORING, REQUEST_APPROVAL, STOP.\n\nAn unrecognised `execution_action` is not permission: fail closed (halt or re-preflight). Well-known code: `not_permission_fail_closed`.\n\nWhen `execution_action` is REQUEST_APPROVAL or STOP, surface the detected patterns and the blast radius, and propose the safer path \u2014 deprecate-then-remove, an additive change, or a new version \u2014 rather than shipping the break.\n\n`CONTINUE_WITH_MONITORING` requires a wired monitoring sink (`monitoringSinkWired`). It is not "proceed with caution" without monitoring.\n\nDo not call CodeRifts tools for a documentation-only change (README, guides, comments) with no contract artifact content change.\n\nIf you already hold a chain receipt and only need authenticity/lifecycle: `verify_receipt`. If you need a past decision by id: `get_decision_details`. Neither replaces preflight for a new change set.\n\nThe CodeRifts MCP server exposes exactly three tools \u2014 `preflight_change_set`, `verify_receipt`, `get_decision_details`. Do not invent or assume others.\n\nA receipt authorizes ONE operation: a merge receipt does not authorize a deploy. Before a different operation (deploy, publish), call `preflight_change_set` with `context.operation` set to that operation \u2014 reusing a differently-scoped receipt is not permitted and will fail at the gate.\n\nA stale or superseded receipt on a changed head requires a NEW preflight \u2014 `verify_receipt` cannot re-diff.\n\nFor mutating tools, put only the guarded version in the agent\'s tool table; keep the raw handler host-only and unreachable from that table. How you name tools is yours \u2014 this is a reachability property, not a product rename of host tools. CodeRifts cannot see or stop a raw call the host makes outside the table it returns; adopt this as a host convention, not as a guarantee from the package.\n\nCodeRifts reports a governance decision and `execution_action`; it does not by itself block merges. Blocking requires separate repository configuration (required status checks, enforcement) that this rule file does not set.\n\nTo act (mutate a contract, merge, deploy, or publish): call `preflight_change_set` with `preflight_mode` authorize. Analyze is informational (`may_execute` is always false) and is not permission. Read `execution_action` on the `decision_result` envelope.\n\nBefore acting under a held receipt: call `verify_receipt` with the intended `context` (operation, environment, repository, branch, pull_request) for THIS attempt. Do not act on a receipt whose scope does not match.\n\nAct only when `currently_authorized` is true (`control_envelope.receipt_view.currently_authorized`). A valid-looking token is not permission if `currently_authorized` is false or omitted.\n\nCommit / CAS evidence is a separate measurement (`commit_observation` on GuardOutcome). It is not a substitute for authorize + `currently_authorized`. Production hosts that want the fail-closed conjunction lock it with `profile: ENFORCING_STRICT` on withCodeRifts.\n\nIf the host requests an execution grant (opt-in `include_execution_grant`), the grant is bound to operation + target + after-payload (`scope_hash`) and is short-lived \u2014 never reuse it after the after-payload changes.\n\nAn ATOMIC-profile grant carries `state_nonce` and is single-use at the executor \u2014 if the executor has consumed the nonce, re-preflight; do not retry the same grant.\n\nWith a proven tenant\u2194repo binding you may request `derivation:"server"` instead of assembling `artifacts[]` yourself (`context.repository` + `context.base` + `context.head` required; caller-supplied artifacts are rejected on that path).\n\nA commit is only proven when an executor attestation verifies (customer-held executor key, `cas_evidence: executor_attested`); otherwise say "authorized, commit not proven".\n',
+      "CLAUDE.md": '<!-- GENERATED from CodeRifts agent-host-rule (single source). Do not edit rule sentences by hand; run: node scripts/generate-agent-host-files.js -->\n\n# CodeRifts\n\nCall `preflight_change_set` before merge, deploy, publish, or tool registration when a contract artifact changed (OpenAPI/Swagger, GraphQL, gRPC/protobuf, AsyncAPI, MCP manifest, or agent tool schemas).\n\nSend the complete base-to-head change set: every changed contract artifact with full before and after content. Do not send a single-file subset when other contract files also change.\n\nBranch on `execution_action` only. Do not branch on `decision` and do not branch on `safe_for_agent` (not_for_control_flow_use_execution_action). Canonical `execution_action` values: CONTINUE, CONTINUE_WITH_MONITORING, REQUEST_APPROVAL, STOP.\n\nAn unrecognised `execution_action` is not permission: fail closed (halt or re-preflight). Well-known code: `not_permission_fail_closed`.\n\nWhen `execution_action` is REQUEST_APPROVAL or STOP, surface the detected patterns and the blast radius, and propose the safer path \u2014 deprecate-then-remove, an additive change, or a new version \u2014 rather than shipping the break.\n\n`CONTINUE_WITH_MONITORING` requires a wired monitoring sink (`monitoringSinkWired`). It is not "proceed with caution" without monitoring.\n\nDo not call CodeRifts tools for a documentation-only change (README, guides, comments) with no contract artifact content change.\n\nIf you already hold a chain receipt and only need authenticity/lifecycle: `verify_receipt`. If you need a past decision by id: `get_decision_details`. Neither replaces preflight for a new change set.\n\nThe CodeRifts MCP server exposes exactly three tools \u2014 `preflight_change_set`, `verify_receipt`, `get_decision_details`. Do not invent or assume others.\n\nA receipt authorizes ONE operation: a merge receipt does not authorize a deploy. Before a different operation (deploy, publish), call `preflight_change_set` with `context.operation` set to that operation \u2014 reusing a differently-scoped receipt is not permitted and will fail at the gate.\n\nA stale or superseded receipt on a changed head requires a NEW preflight \u2014 `verify_receipt` cannot re-diff.\n\nFor mutating tools, put only the guarded version in the agent\'s tool table; keep the raw handler host-only and unreachable from that table. How you name tools is yours \u2014 this is a reachability property, not a product rename of host tools. CodeRifts cannot see or stop a raw call the host makes outside the table it returns; adopt this as a host convention, not as a guarantee from the package.\n\nCodeRifts reports a governance decision and `execution_action`; it does not by itself block merges. Blocking requires separate repository configuration (required status checks, enforcement) that this rule file does not set.\n\nTo act (mutate a contract, merge, deploy, or publish): call `preflight_change_set` with `preflight_mode` authorize. Analyze is informational (`may_execute` is always false) and is not permission. Read `execution_action` on the `decision_result` envelope.\n\nBefore acting under a held receipt: call `verify_receipt` with the intended `context` (operation, environment, repository, branch, pull_request) for THIS attempt. Do not act on a receipt whose scope does not match.\n\nAct only when `currently_authorized` is true (`control_envelope.receipt_view.currently_authorized`). A valid-looking token is not permission if `currently_authorized` is false or omitted.\n\nCommit / CAS evidence is a separate measurement (`commit_observation` on GuardOutcome). It is not a substitute for authorize + `currently_authorized`. Production hosts that want the fail-closed conjunction lock it with `profile: ENFORCING_STRICT` on withCodeRifts.\n\nIf the host requests an execution grant (opt-in `include_execution_grant`), the grant is bound to operation + target + after-payload (`scope_hash`) and is short-lived \u2014 never reuse it after the after-payload changes.\n\nAn ATOMIC-profile grant carries `state_nonce` and is single-use at the executor \u2014 if the executor has consumed the nonce, re-preflight; do not retry the same grant.\n\nWith a proven tenant\u2194repo binding you may request `derivation:"server"` instead of assembling `artifacts[]` yourself (`context.repository` + `context.base` + `context.head` required; caller-supplied artifacts are rejected on that path).\n\nA commit is only proven when an executor attestation verifies (customer-held executor key, `cas_evidence: executor_attested`); otherwise say "authorized, commit not proven".\n',
+      ".cursor/rules/coderifts.mdc": '---\ndescription: CodeRifts API governance \u2014 when to preflight and how to branch\nglobs:\nalwaysApply: true\n---\n\n<!-- GENERATED from CodeRifts agent-host-rule (single source). Do not edit rule sentences by hand; run: node scripts/generate-agent-host-files.js -->\n\n# CodeRifts\n\nCall `preflight_change_set` before merge, deploy, publish, or tool registration when a contract artifact changed (OpenAPI/Swagger, GraphQL, gRPC/protobuf, AsyncAPI, MCP manifest, or agent tool schemas).\n\nSend the complete base-to-head change set: every changed contract artifact with full before and after content. Do not send a single-file subset when other contract files also change.\n\nBranch on `execution_action` only. Do not branch on `decision` and do not branch on `safe_for_agent` (not_for_control_flow_use_execution_action). Canonical `execution_action` values: CONTINUE, CONTINUE_WITH_MONITORING, REQUEST_APPROVAL, STOP.\n\nAn unrecognised `execution_action` is not permission: fail closed (halt or re-preflight). Well-known code: `not_permission_fail_closed`.\n\nWhen `execution_action` is REQUEST_APPROVAL or STOP, surface the detected patterns and the blast radius, and propose the safer path \u2014 deprecate-then-remove, an additive change, or a new version \u2014 rather than shipping the break.\n\n`CONTINUE_WITH_MONITORING` requires a wired monitoring sink (`monitoringSinkWired`). It is not "proceed with caution" without monitoring.\n\nDo not call CodeRifts tools for a documentation-only change (README, guides, comments) with no contract artifact content change.\n\nIf you already hold a chain receipt and only need authenticity/lifecycle: `verify_receipt`. If you need a past decision by id: `get_decision_details`. Neither replaces preflight for a new change set.\n\nThe CodeRifts MCP server exposes exactly three tools \u2014 `preflight_change_set`, `verify_receipt`, `get_decision_details`. Do not invent or assume others.\n\nA receipt authorizes ONE operation: a merge receipt does not authorize a deploy. Before a different operation (deploy, publish), call `preflight_change_set` with `context.operation` set to that operation \u2014 reusing a differently-scoped receipt is not permitted and will fail at the gate.\n\nA stale or superseded receipt on a changed head requires a NEW preflight \u2014 `verify_receipt` cannot re-diff.\n\nFor mutating tools, put only the guarded version in the agent\'s tool table; keep the raw handler host-only and unreachable from that table. How you name tools is yours \u2014 this is a reachability property, not a product rename of host tools. CodeRifts cannot see or stop a raw call the host makes outside the table it returns; adopt this as a host convention, not as a guarantee from the package.\n\nCodeRifts reports a governance decision and `execution_action`; it does not by itself block merges. Blocking requires separate repository configuration (required status checks, enforcement) that this rule file does not set.\n\nTo act (mutate a contract, merge, deploy, or publish): call `preflight_change_set` with `preflight_mode` authorize. Analyze is informational (`may_execute` is always false) and is not permission. Read `execution_action` on the `decision_result` envelope.\n\nBefore acting under a held receipt: call `verify_receipt` with the intended `context` (operation, environment, repository, branch, pull_request) for THIS attempt. Do not act on a receipt whose scope does not match.\n\nAct only when `currently_authorized` is true (`control_envelope.receipt_view.currently_authorized`). A valid-looking token is not permission if `currently_authorized` is false or omitted.\n\nCommit / CAS evidence is a separate measurement (`commit_observation` on GuardOutcome). It is not a substitute for authorize + `currently_authorized`. Production hosts that want the fail-closed conjunction lock it with `profile: ENFORCING_STRICT` on withCodeRifts.\n\nIf the host requests an execution grant (opt-in `include_execution_grant`), the grant is bound to operation + target + after-payload (`scope_hash`) and is short-lived \u2014 never reuse it after the after-payload changes.\n\nAn ATOMIC-profile grant carries `state_nonce` and is single-use at the executor \u2014 if the executor has consumed the nonce, re-preflight; do not retry the same grant.\n\nWith a proven tenant\u2194repo binding you may request `derivation:"server"` instead of assembling `artifacts[]` yourself (`context.repository` + `context.base` + `context.head` required; caller-supplied artifacts are rejected on that path).\n\nA commit is only proven when an executor attestation verifies (customer-held executor key, `cas_evidence: executor_attested`); otherwise say "authorized, commit not proven".\n',
+      ".github/copilot-instructions.md": '<!-- GENERATED from CodeRifts agent-host-rule (single source). Do not edit rule sentences by hand; run: node scripts/generate-agent-host-files.js -->\n\n# CodeRifts instructions for GitHub Copilot\n\nCall `preflight_change_set` before merge, deploy, publish, or tool registration when a contract artifact changed (OpenAPI/Swagger, GraphQL, gRPC/protobuf, AsyncAPI, MCP manifest, or agent tool schemas).\n\nSend the complete base-to-head change set: every changed contract artifact with full before and after content. Do not send a single-file subset when other contract files also change.\n\nBranch on `execution_action` only. Do not branch on `decision` and do not branch on `safe_for_agent` (not_for_control_flow_use_execution_action). Canonical `execution_action` values: CONTINUE, CONTINUE_WITH_MONITORING, REQUEST_APPROVAL, STOP.\n\nAn unrecognised `execution_action` is not permission: fail closed (halt or re-preflight). Well-known code: `not_permission_fail_closed`.\n\nWhen `execution_action` is REQUEST_APPROVAL or STOP, surface the detected patterns and the blast radius, and propose the safer path \u2014 deprecate-then-remove, an additive change, or a new version \u2014 rather than shipping the break.\n\n`CONTINUE_WITH_MONITORING` requires a wired monitoring sink (`monitoringSinkWired`). It is not "proceed with caution" without monitoring.\n\nDo not call CodeRifts tools for a documentation-only change (README, guides, comments) with no contract artifact content change.\n\nIf you already hold a chain receipt and only need authenticity/lifecycle: `verify_receipt`. If you need a past decision by id: `get_decision_details`. Neither replaces preflight for a new change set.\n\nThe CodeRifts MCP server exposes exactly three tools \u2014 `preflight_change_set`, `verify_receipt`, `get_decision_details`. Do not invent or assume others.\n\nA receipt authorizes ONE operation: a merge receipt does not authorize a deploy. Before a different operation (deploy, publish), call `preflight_change_set` with `context.operation` set to that operation \u2014 reusing a differently-scoped receipt is not permitted and will fail at the gate.\n\nA stale or superseded receipt on a changed head requires a NEW preflight \u2014 `verify_receipt` cannot re-diff.\n\nFor mutating tools, put only the guarded version in the agent\'s tool table; keep the raw handler host-only and unreachable from that table. How you name tools is yours \u2014 this is a reachability property, not a product rename of host tools. CodeRifts cannot see or stop a raw call the host makes outside the table it returns; adopt this as a host convention, not as a guarantee from the package.\n\nCodeRifts reports a governance decision and `execution_action`; it does not by itself block merges. Blocking requires separate repository configuration (required status checks, enforcement) that this rule file does not set.\n\nTo act (mutate a contract, merge, deploy, or publish): call `preflight_change_set` with `preflight_mode` authorize. Analyze is informational (`may_execute` is always false) and is not permission. Read `execution_action` on the `decision_result` envelope.\n\nBefore acting under a held receipt: call `verify_receipt` with the intended `context` (operation, environment, repository, branch, pull_request) for THIS attempt. Do not act on a receipt whose scope does not match.\n\nAct only when `currently_authorized` is true (`control_envelope.receipt_view.currently_authorized`). A valid-looking token is not permission if `currently_authorized` is false or omitted.\n\nCommit / CAS evidence is a separate measurement (`commit_observation` on GuardOutcome). It is not a substitute for authorize + `currently_authorized`. Production hosts that want the fail-closed conjunction lock it with `profile: ENFORCING_STRICT` on withCodeRifts.\n\nIf the host requests an execution grant (opt-in `include_execution_grant`), the grant is bound to operation + target + after-payload (`scope_hash`) and is short-lived \u2014 never reuse it after the after-payload changes.\n\nAn ATOMIC-profile grant carries `state_nonce` and is single-use at the executor \u2014 if the executor has consumed the nonce, re-preflight; do not retry the same grant.\n\nWith a proven tenant\u2194repo binding you may request `derivation:"server"` instead of assembling `artifacts[]` yourself (`context.repository` + `context.base` + `context.head` required; caller-supplied artifacts are rejected on that path).\n\nA commit is only proven when an executor attestation verifies (customer-held executor key, `cas_evidence: executor_attested`); otherwise say "authorized, commit not proven".\n',
+      "coderifts-langgraph-policy.js": '// GENERATED from CodeRifts agent-host-rule (single source). Do not edit rule sentences by hand; run: node scripts/generate-agent-host-files.js\n// System policy string for a LangGraph (or similar) agent. Content is generated;\n// identical rule sentences to AGENTS.md / other formats.\n\'use strict\';\n\nmodule.exports = "Call `preflight_change_set` before merge, deploy, publish, or tool registration when a contract artifact changed (OpenAPI/Swagger, GraphQL, gRPC/protobuf, AsyncAPI, MCP manifest, or agent tool schemas).\\n\\nSend the complete base-to-head change set: every changed contract artifact with full before and after content. Do not send a single-file subset when other contract files also change.\\n\\nBranch on `execution_action` only. Do not branch on `decision` and do not branch on `safe_for_agent` (not_for_control_flow_use_execution_action). Canonical `execution_action` values: CONTINUE, CONTINUE_WITH_MONITORING, REQUEST_APPROVAL, STOP.\\n\\nAn unrecognised `execution_action` is not permission: fail closed (halt or re-preflight). Well-known code: `not_permission_fail_closed`.\\n\\nWhen `execution_action` is REQUEST_APPROVAL or STOP, surface the detected patterns and the blast radius, and propose the safer path \u2014 deprecate-then-remove, an additive change, or a new version \u2014 rather than shipping the break.\\n\\n`CONTINUE_WITH_MONITORING` requires a wired monitoring sink (`monitoringSinkWired`). It is not \\"proceed with caution\\" without monitoring.\\n\\nDo not call CodeRifts tools for a documentation-only change (README, guides, comments) with no contract artifact content change.\\n\\nIf you already hold a chain receipt and only need authenticity/lifecycle: `verify_receipt`. If you need a past decision by id: `get_decision_details`. Neither replaces preflight for a new change set.\\n\\nThe CodeRifts MCP server exposes exactly three tools \u2014 `preflight_change_set`, `verify_receipt`, `get_decision_details`. Do not invent or assume others.\\n\\nA receipt authorizes ONE operation: a merge receipt does not authorize a deploy. Before a different operation (deploy, publish), call `preflight_change_set` with `context.operation` set to that operation \u2014 reusing a differently-scoped receipt is not permitted and will fail at the gate.\\n\\nA stale or superseded receipt on a changed head requires a NEW preflight \u2014 `verify_receipt` cannot re-diff.\\n\\nFor mutating tools, put only the guarded version in the agent\'s tool table; keep the raw handler host-only and unreachable from that table. How you name tools is yours \u2014 this is a reachability property, not a product rename of host tools. CodeRifts cannot see or stop a raw call the host makes outside the table it returns; adopt this as a host convention, not as a guarantee from the package.\\n\\nCodeRifts reports a governance decision and `execution_action`; it does not by itself block merges. Blocking requires separate repository configuration (required status checks, enforcement) that this rule file does not set.\\n\\nTo act (mutate a contract, merge, deploy, or publish): call `preflight_change_set` with `preflight_mode` authorize. Analyze is informational (`may_execute` is always false) and is not permission. Read `execution_action` on the `decision_result` envelope.\\n\\nBefore acting under a held receipt: call `verify_receipt` with the intended `context` (operation, environment, repository, branch, pull_request) for THIS attempt. Do not act on a receipt whose scope does not match.\\n\\nAct only when `currently_authorized` is true (`control_envelope.receipt_view.currently_authorized`). A valid-looking token is not permission if `currently_authorized` is false or omitted.\\n\\nCommit / CAS evidence is a separate measurement (`commit_observation` on GuardOutcome). It is not a substitute for authorize + `currently_authorized`. Production hosts that want the fail-closed conjunction lock it with `profile: ENFORCING_STRICT` on withCodeRifts.\\n\\nIf the host requests an execution grant (opt-in `include_execution_grant`), the grant is bound to operation + target + after-payload (`scope_hash`) and is short-lived \u2014 never reuse it after the after-payload changes.\\n\\nAn ATOMIC-profile grant carries `state_nonce` and is single-use at the executor \u2014 if the executor has consumed the nonce, re-preflight; do not retry the same grant.\\n\\nWith a proven tenant\u2194repo binding you may request `derivation:\\"server\\"` instead of assembling `artifacts[]` yourself (`context.repository` + `context.base` + `context.head` required; caller-supplied artifacts are rejected on that path).\\n\\nA commit is only proven when an executor attestation verifies (customer-held executor key, `cas_evidence: executor_attested`); otherwise say \\"authorized, commit not proven\\".";\n',
+      "openai-agent-instructions.md": '<!-- GENERATED from CodeRifts agent-host-rule (single source). Do not edit rule sentences by hand; run: node scripts/generate-agent-host-files.js -->\n\n# CodeRifts agent instructions\n\nCall `preflight_change_set` before merge, deploy, publish, or tool registration when a contract artifact changed (OpenAPI/Swagger, GraphQL, gRPC/protobuf, AsyncAPI, MCP manifest, or agent tool schemas).\n\nSend the complete base-to-head change set: every changed contract artifact with full before and after content. Do not send a single-file subset when other contract files also change.\n\nBranch on `execution_action` only. Do not branch on `decision` and do not branch on `safe_for_agent` (not_for_control_flow_use_execution_action). Canonical `execution_action` values: CONTINUE, CONTINUE_WITH_MONITORING, REQUEST_APPROVAL, STOP.\n\nAn unrecognised `execution_action` is not permission: fail closed (halt or re-preflight). Well-known code: `not_permission_fail_closed`.\n\nWhen `execution_action` is REQUEST_APPROVAL or STOP, surface the detected patterns and the blast radius, and propose the safer path \u2014 deprecate-then-remove, an additive change, or a new version \u2014 rather than shipping the break.\n\n`CONTINUE_WITH_MONITORING` requires a wired monitoring sink (`monitoringSinkWired`). It is not "proceed with caution" without monitoring.\n\nDo not call CodeRifts tools for a documentation-only change (README, guides, comments) with no contract artifact content change.\n\nIf you already hold a chain receipt and only need authenticity/lifecycle: `verify_receipt`. If you need a past decision by id: `get_decision_details`. Neither replaces preflight for a new change set.\n\nThe CodeRifts MCP server exposes exactly three tools \u2014 `preflight_change_set`, `verify_receipt`, `get_decision_details`. Do not invent or assume others.\n\nA receipt authorizes ONE operation: a merge receipt does not authorize a deploy. Before a different operation (deploy, publish), call `preflight_change_set` with `context.operation` set to that operation \u2014 reusing a differently-scoped receipt is not permitted and will fail at the gate.\n\nA stale or superseded receipt on a changed head requires a NEW preflight \u2014 `verify_receipt` cannot re-diff.\n\nFor mutating tools, put only the guarded version in the agent\'s tool table; keep the raw handler host-only and unreachable from that table. How you name tools is yours \u2014 this is a reachability property, not a product rename of host tools. CodeRifts cannot see or stop a raw call the host makes outside the table it returns; adopt this as a host convention, not as a guarantee from the package.\n\nCodeRifts reports a governance decision and `execution_action`; it does not by itself block merges. Blocking requires separate repository configuration (required status checks, enforcement) that this rule file does not set.\n\nTo act (mutate a contract, merge, deploy, or publish): call `preflight_change_set` with `preflight_mode` authorize. Analyze is informational (`may_execute` is always false) and is not permission. Read `execution_action` on the `decision_result` envelope.\n\nBefore acting under a held receipt: call `verify_receipt` with the intended `context` (operation, environment, repository, branch, pull_request) for THIS attempt. Do not act on a receipt whose scope does not match.\n\nAct only when `currently_authorized` is true (`control_envelope.receipt_view.currently_authorized`). A valid-looking token is not permission if `currently_authorized` is false or omitted.\n\nCommit / CAS evidence is a separate measurement (`commit_observation` on GuardOutcome). It is not a substitute for authorize + `currently_authorized`. Production hosts that want the fail-closed conjunction lock it with `profile: ENFORCING_STRICT` on withCodeRifts.\n\nIf the host requests an execution grant (opt-in `include_execution_grant`), the grant is bound to operation + target + after-payload (`scope_hash`) and is short-lived \u2014 never reuse it after the after-payload changes.\n\nAn ATOMIC-profile grant carries `state_nonce` and is single-use at the executor \u2014 if the executor has consumed the nonce, re-preflight; do not retry the same grant.\n\nWith a proven tenant\u2194repo binding you may request `derivation:"server"` instead of assembling `artifacts[]` yourself (`context.repository` + `context.base` + `context.head` required; caller-supplied artifacts are rejected on that path).\n\nA commit is only proven when an executor attestation verifies (customer-held executor key, `cas_evidence: executor_attested`); otherwise say "authorized, commit not proven".\n',
+      "skills/api-governance/SKILL.md": '---\nname: api-governance\ndescription: Before merging or shipping any API or tool-contract change: preflight the change set, then branch on execution_action.\n---\n\n# CodeRifts API Governance\n\n<!-- GENERATED from CodeRifts agent-host-rule (single source). Do not edit rule sentences by hand; run: node scripts/generate-agent-host-files.js -->\n\nCall `preflight_change_set` before merge, deploy, publish, or tool registration when a contract artifact changed (OpenAPI/Swagger, GraphQL, gRPC/protobuf, AsyncAPI, MCP manifest, or agent tool schemas).\n\nSend the complete base-to-head change set: every changed contract artifact with full before and after content. Do not send a single-file subset when other contract files also change.\n\nBranch on `execution_action` only. Do not branch on `decision` and do not branch on `safe_for_agent` (not_for_control_flow_use_execution_action). Canonical `execution_action` values: CONTINUE, CONTINUE_WITH_MONITORING, REQUEST_APPROVAL, STOP.\n\nAn unrecognised `execution_action` is not permission: fail closed (halt or re-preflight). Well-known code: `not_permission_fail_closed`.\n\nWhen `execution_action` is REQUEST_APPROVAL or STOP, surface the detected patterns and the blast radius, and propose the safer path \u2014 deprecate-then-remove, an additive change, or a new version \u2014 rather than shipping the break.\n\n`CONTINUE_WITH_MONITORING` requires a wired monitoring sink (`monitoringSinkWired`). It is not "proceed with caution" without monitoring.\n\nDo not call CodeRifts tools for a documentation-only change (README, guides, comments) with no contract artifact content change.\n\nIf you already hold a chain receipt and only need authenticity/lifecycle: `verify_receipt`. If you need a past decision by id: `get_decision_details`. Neither replaces preflight for a new change set.\n\nThe CodeRifts MCP server exposes exactly three tools \u2014 `preflight_change_set`, `verify_receipt`, `get_decision_details`. Do not invent or assume others.\n\nA receipt authorizes ONE operation: a merge receipt does not authorize a deploy. Before a different operation (deploy, publish), call `preflight_change_set` with `context.operation` set to that operation \u2014 reusing a differently-scoped receipt is not permitted and will fail at the gate.\n\nA stale or superseded receipt on a changed head requires a NEW preflight \u2014 `verify_receipt` cannot re-diff.\n\nFor mutating tools, put only the guarded version in the agent\'s tool table; keep the raw handler host-only and unreachable from that table. How you name tools is yours \u2014 this is a reachability property, not a product rename of host tools. CodeRifts cannot see or stop a raw call the host makes outside the table it returns; adopt this as a host convention, not as a guarantee from the package.\n\nCodeRifts reports a governance decision and `execution_action`; it does not by itself block merges. Blocking requires separate repository configuration (required status checks, enforcement) that this rule file does not set.\n\nTo act (mutate a contract, merge, deploy, or publish): call `preflight_change_set` with `preflight_mode` authorize. Analyze is informational (`may_execute` is always false) and is not permission. Read `execution_action` on the `decision_result` envelope.\n\nBefore acting under a held receipt: call `verify_receipt` with the intended `context` (operation, environment, repository, branch, pull_request) for THIS attempt. Do not act on a receipt whose scope does not match.\n\nAct only when `currently_authorized` is true (`control_envelope.receipt_view.currently_authorized`). A valid-looking token is not permission if `currently_authorized` is false or omitted.\n\nCommit / CAS evidence is a separate measurement (`commit_observation` on GuardOutcome). It is not a substitute for authorize + `currently_authorized`. Production hosts that want the fail-closed conjunction lock it with `profile: ENFORCING_STRICT` on withCodeRifts.\n\nIf the host requests an execution grant (opt-in `include_execution_grant`), the grant is bound to operation + target + after-payload (`scope_hash`) and is short-lived \u2014 never reuse it after the after-payload changes.\n\nAn ATOMIC-profile grant carries `state_nonce` and is single-use at the executor \u2014 if the executor has consumed the nonce, re-preflight; do not retry the same grant.\n\nWith a proven tenant\u2194repo binding you may request `derivation:"server"` instead of assembling `artifacts[]` yourself (`context.repository` + `context.base` + `context.head` required; caller-supplied artifacts are rejected on that path).\n\nA commit is only proven when an executor attestation verifies (customer-held executor key, `cas_evidence: executor_attested`); otherwise say "authorized, commit not proven".\n',
       ".claude/settings.json": '{\n  "hooks": {\n    "PreToolUse": [\n      {\n        "matcher": "Write|Edit|MultiEdit",\n        "hooks": [\n          {\n            "type": "command",\n            "command": "coderifts claude-hook",\n            "timeout": 60\n          }\n        ]\n      }\n    ]\n  }\n}\n',
       ".cursor/hooks.json": '{\n  "version": 1,\n  "hooks": {\n    "preToolUse": [\n      {\n        "command": "coderifts cursor-hook",\n        "matcher": "Write|Delete",\n        "timeout": 60,\n        "failClosed": true\n      }\n    ]\n  }\n}\n'
     });
@@ -210222,6 +210224,46 @@ jobs:
           # monitoring-attestation: \${{ vars.CODERIFTS_MONITORING_ATTESTATION }}
           # monitoring-keyring: \${{ github.workspace }}/.coderifts/monitoring-keys.json
 `;
+    var STRICT_WORKFLOW_BODY = `# Written by \`coderifts init --agents --strict\`. Re-run is a no-op when this file already exists.
+#
+# STRICT template: require-verified-monitoring is ON.
+# This workflow POSTS the Check Run named exactly:
+#   CodeRifts / contract-gate
+# It does NOT enable branch protection and it does NOT make that check required.
+# Turn the check on in repo settings to enforce (${ENFORCEMENT_DOC}).
+#
+# Copied from the gate repo example (examples/contract-gate.yml) \u2014 action at @v0.
+name: CodeRifts Contract Gate
+
+# Run on EVERY pull request. Do NOT add paths: filters \u2014 a required check that
+# never runs leaves the PR mergeable (see ENFORCEMENT.md, "fail-closed on absence").
+on:
+  pull_request:
+
+permissions:
+  checks: write
+  contents: read
+
+jobs:
+  contract-gate:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          # REQUIRED: the gate derives the change set from git diff base...head.
+          fetch-depth: 0
+
+      - uses: ${GATE_ACTION}
+        with:
+          api-key: \${{ secrets.CODERIFTS_API_KEY }}
+          # api-url: https://app.coderifts.com   # override only for self-hosted
+          require-verified-monitoring: 'true'
+          monitoring-attestation: \${{ vars.CODERIFTS_MONITORING_ATTESTATION }}
+          monitoring-keyring: \${{ github.workspace }}/.coderifts/monitoring-keys.json
+`;
+    function workflowBody(strict) {
+      return strict ? STRICT_WORKFLOW_BODY : WORKFLOW_BODY;
+    }
     function writeContractGateWorkflow(opts = {}) {
       const outDir = opts.outDir;
       const dest = path.join(outDir, WORKFLOW_REL);
@@ -210232,6 +210274,7 @@ jobs:
         fs.writeFileSync(p, c, "utf8");
       });
       const dryRun = !!opts.dryRun;
+      const body = workflowBody(!!opts.strict);
       if (exists(dest)) {
         let onDisk = "";
         try {
@@ -210239,13 +210282,13 @@ jobs:
         } catch {
           onDisk = "";
         }
-        const reason = onDisk === WORKFLOW_BODY ? "already present, identical" : "exists; installer does not clobber workflows";
+        const reason = onDisk === body ? "already present, identical" : "exists; installer does not clobber workflows";
         return { action: "skip", relPath: WORKFLOW_REL, reason, exitCode: 0 };
       }
       if (dryRun) {
         return { action: "create", relPath: WORKFLOW_REL, reason: "created", exitCode: 0, dryRun: true };
       }
-      writeFile(dest, WORKFLOW_BODY);
+      writeFile(dest, body);
       return { action: "create", relPath: WORKFLOW_REL, reason: "created", exitCode: 0 };
     }
     function workflowPresent(outDir, deps = {}) {
@@ -210262,6 +210305,7 @@ jobs:
     module2.exports = {
       WORKFLOW_REL,
       WORKFLOW_BODY,
+      STRICT_WORKFLOW_BODY,
       GATE_ACTION,
       ENFORCEMENT_DOC,
       writeContractGateWorkflow,
@@ -210293,7 +210337,7 @@ var require_init_agents = __commonJS({
       workflowPresent
     } = require_contract_gate_workflow();
     if (process.env.NO_COLOR) chalk.level = 0;
-    var USAGE = `Usage: coderifts init --agents [--hosts=claude,cursor,copilot,all] [--no-hook] [--no-workflow] [--dry-run] [--check] [--out <dir>]
+    var USAGE = `Usage: coderifts init --agents [--hosts=claude,cursor,copilot,all] [--no-hook] [--no-workflow] [--strict] [--dry-run] [--check] [--out <dir>]
 
   Wire this repository for governed agent work in ONE command:
     1. repo-level MCP config (per host: .mcp.json / .cursor/mcp.json / .vscode/mcp.json)
@@ -210304,6 +210348,7 @@ var require_init_agents = __commonJS({
   --hosts <list>   claude, cursor, copilot, or all (default: all)
   --no-hook        Skip host hook install
   --no-workflow    Skip writing the CI workflow
+  --strict         STRICT workflow: require-verified-monitoring: true (default template stays false)
   --dry-run        Print the plan without writing
   --check          Report which of the four pieces are present (no write)
   --out <dir>      Target directory (default: cwd)
@@ -210363,9 +210408,9 @@ ${USAGE}`
       return { hosts, allRules: false };
     }
     function agentFlagsRequireAgents(options = {}) {
-      return !!(options.check || options.dryRun || options.out || options.hosts != null && String(options.hosts) !== "all" || options.hook === false || options.workflow === false || options.noHook === true || options.noWorkflow === true);
+      return !!(options.check || options.dryRun || options.out || options.hosts != null && String(options.hosts) !== "all" || options.hook === false || options.workflow === false || options.noHook === true || options.noWorkflow === true || options.strict === true);
     }
-    var AGENT_FLAGS_REQUIRE_AGENTS = "Error: --check/--dry-run/--out/--hosts/--no-hook/--no-workflow require --agents";
+    var AGENT_FLAGS_REQUIRE_AGENTS = "Error: --check/--dry-run/--out/--hosts/--no-hook/--no-workflow/--strict require --agents";
     var AGENTS_NO_TEMPLATE = "Error: init --agents does not take a template. Did you mean `coderifts init --agents` (not `coderifts init agent`)?";
     function rulePathsForHosts(hosts, allRules) {
       const out = [SHARED_RULE];
@@ -210628,12 +210673,12 @@ ${USAGE}`
       if (!wantWorkflow(options)) {
         items.push({ section: "workflow", action: "skip", relPath: WORKFLOW_REL, note: "--no-workflow" });
       } else {
-        const r = writeContractGateWorkflow({ outDir, dryRun, ...fsDeps });
+        const r = writeContractGateWorkflow({ outDir, dryRun, strict: !!options.strict, ...fsDeps });
         items.push({
           section: "workflow",
           action: r.action,
           relPath: r.relPath,
-          note: r.action === "create" ? `${GATE_ACTION}` : r.reason
+          note: r.action === "create" ? options.strict ? `${GATE_ACTION} (STRICT require-verified-monitoring)` : GATE_ACTION : r.reason
         });
       }
       const created = items.filter((i) => i.action === "create").length;
@@ -236619,22 +236664,23 @@ var require_setup_required_check = __commonJS({
         observation_error: null
       };
     }
+    var ENFORCING_ISSUER_APP_ID = 15368;
     function buildProtectionUpdatePayload(protection, contextName = CHECK_NAME) {
       const p = protection && typeof protection === "object" ? protection : {};
       const prev = p.required_status_checks && typeof p.required_status_checks === "object" ? p.required_status_checks : {};
       const contexts = extractRequiredContexts(p);
       const nextContexts = contexts.includes(contextName) ? contexts.slice() : contexts.concat([contextName]);
+      const checksOut = nextContexts.map((ctx) => {
+        const existing = Array.isArray(prev.checks) ? prev.checks.find((c) => c && c.context === ctx) : null;
+        if (existing && existing.app_id != null) return { context: ctx, app_id: existing.app_id };
+        if (ctx === contextName) return { context: ctx, app_id: ENFORCING_ISSUER_APP_ID };
+        return { context: ctx };
+      });
       const body = {
         required_status_checks: {
           strict: prev.strict === true,
           contexts: nextContexts,
-          // Prefer also sending checks[] when the API used that shape so app ids survive when present.
-          ...Array.isArray(prev.checks) && prev.checks.length ? {
-            checks: nextContexts.map((ctx) => {
-              const existing = prev.checks.find((c) => c && c.context === ctx);
-              return existing && existing.app_id != null ? { context: ctx, app_id: existing.app_id } : { context: ctx };
-            })
-          } : {}
+          checks: checksOut
         },
         enforce_admins: !!(p.enforce_admins && p.enforce_admins.enabled),
         required_pull_request_reviews: p.required_pull_request_reviews ? serializePrReviews(p.required_pull_request_reviews) : null,
@@ -237039,6 +237085,13 @@ var require_setup_required_check = __commonJS({
         log(chalk.green(`Success: '${CHECK_NAME}' is now required.`));
         log(`  repo:   ${owner}/${repo}`);
         log(`  branch: ${branch}`);
+        log(`  issuer: app_id ${ENFORCING_ISSUER_APP_ID} (github-actions) \u2014 the contract-gate Action`);
+        log("");
+        log(chalk.yellow("  What this does NOT achieve:"));
+        log("  \xB7 The issuer is shared by every GitHub Actions workflow in this repository, so anyone");
+        log("    who can edit .github/workflows can post a passing check under it. Govern those paths.");
+        log("  \xB7 A repository admin with enforce_admins disabled still merges past a required check.");
+        log("  \xB7 This binds the merge gate only. Deploy and publish are separate paths.");
       }
       return finish(EXIT.OK, {
         ...basePayload,
@@ -237059,6 +237112,7 @@ var require_setup_required_check = __commonJS({
       HINT_403_PERMISSION,
       HINT_403_BOTH,
       buildProtectionUpdatePayload,
+      ENFORCING_ISSUER_APP_ID,
       detectRulesetRequiredCheck,
       parseGitHubRemote,
       resolveOwnerRepo,
@@ -237401,12 +237455,17 @@ var require_github_enforcement = __commonJS({
     var EVIDENCE_SPEC = "provider-enforcement-evidence.v2";
     var CODERIFTS_GITHUB_APP_SLUG = "coderifts";
     var CODERIFTS_GITHUB_APP_ID = 2860592;
+    var GITHUB_ACTIONS_APP_ID = 15368;
+    var GITHUB_ACTIONS_APP_SLUG = "github-actions";
+    var ENFORCING_ISSUER_APP_ID = GITHUB_ACTIONS_APP_ID;
     var REASON = Object.freeze({
       ISSUER_BOUND: "issuer_bound",
       NO_ISSUER_BINDING: "no_issuer_binding",
       LEGACY_CONTEXTS_NO_ISSUER_BINDING: "legacy_contexts_no_issuer_binding",
       ISSUER_APP_ID_MISMATCH: "issuer_app_id_mismatch",
       ISSUER_APP_ID_UNRESOLVABLE: "issuer_app_id_unresolvable",
+      /** Bound to the CodeRifts App check, whose conclusion depends on a server setting we cannot read. */
+      ISSUER_ADVISORY_APP_UNOBSERVABLE: "issuer_advisory_app_conclusion_unobservable",
       REQUIRED_CHECK_ABSENT: "required_check_absent",
       ACTIVE_JOB: "active_job",
       NO_ACTIVE_JOB: "no_active_job",
@@ -237452,14 +237511,25 @@ var require_github_enforcement = __commonJS({
       }
       const ours = ourCheckEntries(rsc);
       if (ours.length > 0) {
-        const bound = ours.find((c) => numericAppId(c.app_id) === expected);
-        if (bound) {
+        const enforcing = ours.find((c) => numericAppId(c.app_id) === ENFORCING_ISSUER_APP_ID);
+        if (enforcing) {
           return {
             status: STATUS.VERIFIED,
             reason_code: REASON.ISSUER_BOUND,
-            reason: `required_status_checks.checks binds "${CHECK_NAME}" to GitHub App id ${expected} (slug ${CODERIFTS_GITHUB_APP_SLUG})`,
+            reason: `required_status_checks.checks binds "${CHECK_NAME}" to GitHub App id ${ENFORCING_ISSUER_APP_ID} (slug ${GITHUB_ACTIONS_APP_SLUG}) \u2014 the issuer the coderifts/contract-gate Action posts under, which concludes failure and blocks the merge. RESIDUAL: that issuer is shared by every GitHub Actions workflow in this repository, so anyone who can edit .github/workflows can post a passing check under it. Govern those paths too.`,
+            required_contexts: contexts,
+            app_id: ENFORCING_ISSUER_APP_ID
+          };
+        }
+        const advisory = ours.find((c) => numericAppId(c.app_id) === expected);
+        if (advisory) {
+          return {
+            status: STATUS.UNVERIFIABLE,
+            reason_code: REASON.ISSUER_ADVISORY_APP_UNOBSERVABLE,
+            reason: `required_status_checks.checks binds "${CHECK_NAME}" to CodeRifts App id ${expected} (slug ${CODERIFTS_GITHUB_APP_SLUG}). That check is advisory by default: it concludes "neutral" unless the server runs with MERGEGATE_ENFORCE=true, and GitHub treats a neutral required check as passing. MERGEGATE_ENFORCE is a server-side setting this CLI cannot observe, so whether this pin gates CANNOT BE DETERMINED FROM HERE. To get a check that provably blocks, bind "${CHECK_NAME}" to app_id ${ENFORCING_ISSUER_APP_ID} (${GITHUB_ACTIONS_APP_SLUG}) \u2014 the contract-gate Action.`,
             required_contexts: contexts,
-            app_id: expected
+            app_id: expected,
+            expected_app_id: ENFORCING_ISSUER_APP_ID
           };
         }
         const other = ours.find((c) => numericAppId(c.app_id) != null);
@@ -237467,10 +237537,10 @@ var require_github_enforcement = __commonJS({
           return {
             status: STATUS.NOT_VERIFIED,
             reason_code: REASON.ISSUER_APP_ID_MISMATCH,
-            reason: `required_status_checks.checks binds "${CHECK_NAME}" to app_id ${numericAppId(other.app_id)}, not CodeRifts App id ${expected}`,
+            reason: `required_status_checks.checks binds "${CHECK_NAME}" to app_id ${numericAppId(other.app_id)}, not the contract-gate Action issuer ${ENFORCING_ISSUER_APP_ID} (${GITHUB_ACTIONS_APP_SLUG})`,
             required_contexts: contexts,
             app_id: numericAppId(other.app_id),
-            expected_app_id: expected
+            expected_app_id: ENFORCING_ISSUER_APP_ID
           };
         }
         return {
@@ -237857,6 +237927,8 @@ var require_github_enforcement = __commonJS({
       GATE_ACTION,
       CODERIFTS_GITHUB_APP_ID,
       CODERIFTS_GITHUB_APP_SLUG,
+      ENFORCING_ISSUER_APP_ID,
+      GITHUB_ACTIONS_APP_SLUG,
       resolveGitHubAuth,
       redactSecrets,
       createGitHubHttpClient,
@@ -241087,7 +241159,7 @@ program.command("registry-gate [dir]").description("Admit a directory of OpenAPI
   process.exitCode = code;
   process.exit(code);
 });
-program.command("init [template]").description("Generate a .coderifts.yml from a policy template, or wire governed agent work (--agents)").option("--agents", "Wire this repo for governed agent work (MCP config, rule files, host hook, CI gate)").option("--hosts <list>", "With --agents: claude, cursor, copilot, or all (default: all)", "all").option("--no-hook", "With --agents: skip host hook install").option("--no-workflow", "With --agents: skip writing .github/workflows/coderifts.yml").option("--dry-run", "With --agents: print the plan without writing").option("--check", "With --agents: report which of the four pieces are present (no write)").option("--out <dir>", "With --agents: target directory (default: cwd)").action(async (template, options) => {
+program.command("init [template]").description("Generate a .coderifts.yml from a policy template, or wire governed agent work (--agents)").option("--agents", "Wire this repo for governed agent work (MCP config, rule files, host hook, CI gate)").option("--hosts <list>", "With --agents: claude, cursor, copilot, or all (default: all)", "all").option("--no-hook", "With --agents: skip host hook install").option("--no-workflow", "With --agents: skip writing .github/workflows/coderifts.yml").option("--strict", "With --agents: STRICT workflow (require-verified-monitoring: true)").option("--dry-run", "With --agents: print the plan without writing").option("--check", "With --agents: report which of the four pieces are present (no write)").option("--out <dir>", "With --agents: target directory (default: cwd)").action(async (template, options) => {
   const {
     runInitAgents,
     agentFlagsRequireAgents,
@@ -241106,7 +241178,8 @@ program.command("init [template]").description("Generate a .coderifts.yml from a
       workflow: options.workflow,
       dryRun: !!options.dryRun,
       check: !!options.check,
-      out: options.out
+      out: options.out,
+      strict: !!options.strict
     });
     if (result && typeof result.exitCode === "number") {
       process.exitCode = result.exitCode;

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "coderifts",
-  "version": "4.9.0",
+  "version": "4.11.0",
   "description": "Detect breaking API changes from the command line. Works locally or with the CodeRifts cloud API.",
   "author": "CodeRifts <hello@coderifts.com>",
   "license": "MIT",