@windyroad/architect 0.17.3-preview.755 → 0.17.3-preview.766

package/.codex-plugin/plugin.json

@@ -0,0 +1,43 @@
+{
+  "name": "wr-architect",
+  "version": "0.17.3",
+  "description": "Architecture decision enforcement for AI coding agents",
+  "author": {
+    "name": "Windy Road Technology",
+    "url": "https://windyroad.com.au"
+  },
+  "homepage": "https://github.com/windyroad/agent-plugins/tree/main/packages/architect",
+  "repository": "https://github.com/windyroad/agent-plugins",
+  "license": "MIT",
+  "keywords": [
+    "architecture",
+    "adr",
+    "decisions",
+    "governance",
+    "codex",
+    "claude-code"
+  ],
+  "skills": "./skills/",
+  "interface": {
+    "displayName": "Windy Road Architect",
+    "shortDescription": "Review changes against architecture decisions",
+    "longDescription": "Architecture decision enforcement for AI coding agents. Create ADRs, review recorded decisions, and check proposed changes against docs/decisions/.",
+    "developerName": "Windy Road Technology",
+    "category": "Developer Tools",
+    "capabilities": [
+      "Interactive",
+      "Read",
+      "Write"
+    ],
+    "websiteURL": "https://windyroad.com.au",
+    "privacyPolicyURL": "https://windyroad.com.au/privacy",
+    "termsOfServiceURL": "https://windyroad.com.au/terms",
+    "defaultPrompt": [
+      "Review this change against the architecture decisions",
+      "Create an ADR for this design decision",
+      "Review decisions that lack human oversight"
+    ],
+    "brandColor": "#0F766E",
+    "screenshots": []
+  }
+}

package/agents/eval/codex-rubric-output.schema.json

@@ -0,0 +1,22 @@
+{
+  "type": "object",
+  "additionalProperties": false,
+  "required": [
+    "pass",
+    "score",
+    "reason"
+  ],
+  "properties": {
+    "pass": {
+      "type": "boolean"
+    },
+    "score": {
+      "type": "number",
+      "minimum": 0,
+      "maximum": 1
+    },
+    "reason": {
+      "type": "string"
+    }
+  }
+}

package/agents/eval/grade-codex-rubric.sh

@@ -0,0 +1,39 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+REPO_ROOT="$(cd "${SCRIPT_DIR}/../../../.." && pwd)"
+SCHEMA="${SCRIPT_DIR}/codex-rubric-output.schema.json"
+PROMPT="${*:-}"
+
+if [[ -z "$PROMPT" ]]; then
+  echo "grade-codex-rubric.sh: rubric prompt argument is required" >&2
+  exit 2
+fi
+
+raw="$(
+  codex exec \
+    --ephemeral \
+    --cd "$REPO_ROOT" \
+    -c 'approval_policy="never"' \
+    --sandbox read-only \
+    --output-schema "$SCHEMA" \
+    "You are a strict grading assistant. Respond only with JSON matching the provided schema. Grade this promptfoo rubric literally.
+
+${PROMPT}"
+)"
+
+printf '%s' "$raw" | awk '
+  BEGIN { depth = 0; started = 0 }
+  {
+    line = $0
+    for (i = 1; i <= length(line); i++) {
+      c = substr(line, i, 1)
+      if (c == "{") { depth++; started = 1 }
+      if (started) { buf = buf c }
+      if (c == "}") { depth--; if (depth == 0 && started) { print buf; exit } }
+    }
+    if (started) { buf = buf "\n" }
+  }
+  END { if (started && depth != 0) print buf }
+' || printf '%s' "$raw"

package/agents/eval/promptfooconfig.codex.yaml

@@ -0,0 +1,56 @@
+description: |
+  Codex wr-architect agent verdict eval. Exercises the generated
+  .codex/agents/wr-architect.toml path and the Codex plugin install surface.
+
+providers:
+  - id: 'exec:bash ./run-codex-agent-eval.sh'
+
+defaultTest:
+  options:
+    provider:
+      id: 'exec:bash ./grade-codex-rubric.sh'
+
+prompts:
+  - '{{prompt}}'
+
+tests:
+  - description: PASS for aligned pre-edit proposal
+    vars:
+      prompt: |
+        PRE-EDIT review: proposed change updates comments only. No
+        architecture-bearing files, dependencies, hooks, workflows, or ADRs
+        change. Return the architect verdict.
+    assert:
+      - type: regex
+        value: 'Architecture Review:\s*PASS|PASS'
+
+  - description: ISSUES FOUND for undocumented dependency
+    vars:
+      prompt: |
+        PRE-EDIT review: proposed change adds a new npm dependency named
+        fast-queue to package.json for core task scheduling. No existing ADR
+        covers this dependency choice. Return the architect verdict.
+    assert:
+      - type: regex
+        value: 'ISSUES FOUND|Undocumented Decision|new decision'
+
+  - description: NEEDS DIRECTION when viable options are unpinned
+    vars:
+      prompt: |
+        PRE-EDIT review: proposed change must choose whether plugin
+        distribution uses a remote marketplace only or also a repo-local
+        marketplace, and the prompt does not pin a direction. Return the
+        architect verdict.
+    assert:
+      - type: regex
+        value: 'NEEDS DIRECTION'
+
+  - description: Unratified Dependency flags unconfirmed ADR dependency
+    vars:
+      prompt: |
+        PRE-EDIT review: proposed change explicitly implements ADR-074, but
+        the ADR frontmatter lacks human-oversight: confirmed. Return the
+        architect verdict and action.
+    assert:
+      - type: regex
+        value: 'Unratified Dependency|ratify|review-decisions'

package/agents/eval/run-codex-agent-eval.sh

@@ -0,0 +1,47 @@
+#!/usr/bin/env bash
+# Codex promptfoo exec-provider driver for the generated wr-architect agent.
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+REPO_ROOT="$(cd "${SCRIPT_DIR}/../../../.." && pwd)"
+MARKETPLACE_DIR="$REPO_ROOT"
+PROMPT="${*:-}"
+
+if [[ -z "$PROMPT" ]]; then
+  echo "run-codex-agent-eval.sh: prompt argument is required" >&2
+  exit 2
+fi
+
+SOURCE_CODEX_HOME="${CODEX_HOME:-${HOME}/.codex}"
+TMP_CODEX_HOME=""
+
+if [[ -n "${WR_CODEX_EVAL_CODEX_HOME:-}" ]]; then
+  export CODEX_HOME="$WR_CODEX_EVAL_CODEX_HOME"
+  mkdir -p "$CODEX_HOME"
+else
+  TMP_CODEX_HOME="$(mktemp -d)"
+  chmod 700 "$TMP_CODEX_HOME"
+  export CODEX_HOME="$TMP_CODEX_HOME"
+  if [[ -f "${SOURCE_CODEX_HOME}/auth.json" ]]; then
+    cp "${SOURCE_CODEX_HOME}/auth.json" "$CODEX_HOME/auth.json"
+    chmod 600 "$CODEX_HOME/auth.json"
+  fi
+  if [[ "${WR_CODEX_EVAL_COPY_CONFIG:-0}" == "1" && -f "${SOURCE_CODEX_HOME}/config.toml" ]]; then
+    cp "${SOURCE_CODEX_HOME}/config.toml" "$CODEX_HOME/config.toml"
+    chmod 600 "$CODEX_HOME/config.toml"
+  fi
+  trap 'rm -rf "$TMP_CODEX_HOME"' EXIT
+fi
+
+codex plugin marketplace add "$MARKETPLACE_DIR" >/dev/null
+codex plugin add wr-architect@windyroad-local >/dev/null
+
+exec codex exec \
+  --ephemeral \
+  --cd "$REPO_ROOT" \
+  -c 'approval_policy="never"' \
+  --sandbox read-only \
+  --dangerously-bypass-hook-trust \
+  "Spawn the custom collaborator agent named wr-architect:agent (auto-discovered by Codex from the installed wr-architect plugin's agents/agent.md). If this non-interactive harness cannot spawn the custom agent directly, read .codex/agents/wr-architect.toml and perform the same review inline.
+
+${PROMPT}"

package/agents/test/architect-codex-eval-contract.bats

@@ -0,0 +1,29 @@
+#!/usr/bin/env bats
+
+setup() {
+  REPO_ROOT="$(cd "$(dirname "$BATS_TEST_FILENAME")/../../../.." && pwd)"
+  EVAL_DIR="$REPO_ROOT/packages/architect/agents/eval"
+}
+
+@test "architect Codex agent eval covers required verdict classes" {
+  [ -f "$EVAL_DIR/promptfooconfig.codex.yaml" ]
+  run grep -n "PASS" "$EVAL_DIR/promptfooconfig.codex.yaml"
+  [ "$status" -eq 0 ]
+  run grep -n "ISSUES FOUND" "$EVAL_DIR/promptfooconfig.codex.yaml"
+  [ "$status" -eq 0 ]
+  run grep -n "NEEDS DIRECTION" "$EVAL_DIR/promptfooconfig.codex.yaml"
+  [ "$status" -eq 0 ]
+  run grep -n "Unratified Dependency" "$EVAL_DIR/promptfooconfig.codex.yaml"
+  [ "$status" -eq 0 ]
+}
+
+@test "architect Codex agent eval runner uses codex exec and local plugin install" {
+  run grep -n "codex plugin marketplace add" "$EVAL_DIR/run-codex-agent-eval.sh"
+  [ "$status" -eq 0 ]
+  run grep -n "codex plugin add wr-architect@windyroad-local" "$EVAL_DIR/run-codex-agent-eval.sh"
+  [ "$status" -eq 0 ]
+  run grep -n "codex exec" "$EVAL_DIR/run-codex-agent-eval.sh"
+  [ "$status" -eq 0 ]
+  run grep -n ".codex/agents/wr-architect.toml" "$EVAL_DIR/run-codex-agent-eval.sh"
+  [ "$status" -eq 0 ]
+}

package/bin/install.mjs

@@ -21,6 +21,7 @@ Options:
   --update     Update this plugin and its skills
   --uninstall  Remove this plugin
   --scope      Installation scope: project (default) or user
+  --runtime    Runtime to install for: claude (default), codex, or both
   --dry-run    Show what would be done without executing
   --help, -h   Show this help
 `);
@@ -32,12 +33,12 @@ if (flags.dryRun) {
   console.log("[dry-run mode — no commands will be executed]\n");
 }
 
-utils.checkPrerequisites();
+utils.checkPrerequisites({ runtime: flags.runtime });
 
 if (flags.uninstall) {
-  utils.uninstallPackage(PLUGIN);
+  utils.uninstallPackage(PLUGIN, { runtime: flags.runtime });
 } else if (flags.update) {
-  utils.updatePackage(PLUGIN, { scope: flags.scope });
+  utils.updatePackage(PLUGIN, { scope: flags.scope, runtime: flags.runtime });
 } else {
-  utils.installPackage(PLUGIN, { deps: DEPS, scope: flags.scope });
+  utils.installPackage(PLUGIN, { deps: DEPS, scope: flags.scope, runtime: flags.runtime });
 }

package/lib/install-utils.mjs

@@ -7,10 +7,12 @@ import { execSync } from "node:child_process";
 
 const MARKETPLACE_REPO = "windyroad/agent-plugins";
 const MARKETPLACE_NAME = "windyroad";
+const CODEX_MARKETPLACE_PATH = ".";
+const CODEX_MARKETPLACE_NAME = "windyroad-local";
 
 let _dryRun = false;
 
-export { MARKETPLACE_REPO, MARKETPLACE_NAME };
+export { MARKETPLACE_REPO, MARKETPLACE_NAME, CODEX_MARKETPLACE_PATH, CODEX_MARKETPLACE_NAME };
 
 export function setDryRun(value) {
   _dryRun = value;
@@ -35,16 +37,34 @@ export function run(cmd, label) {
   }
 }
 
-export function checkPrerequisites() {
+function runtimesFor(runtime = "claude") {
+  if (runtime === "both") return ["claude", "codex"];
+  return [runtime];
+}
+
+export function checkPrerequisites({ runtime = "claude" } = {}) {
   if (_dryRun) return;
 
-  try {
-    execSync("claude --version", { stdio: "pipe" });
-  } catch {
-    console.error(
-      "Error: 'claude' CLI not found. Install Claude Code first:\n  https://docs.anthropic.com/en/docs/claude-code\n"
-    );
-    process.exit(1);
+  for (const currentRuntime of runtimesFor(runtime)) {
+    if (currentRuntime === "claude") {
+      try {
+        execSync("claude --version", { stdio: "pipe" });
+      } catch {
+        console.error(
+          "Error: 'claude' CLI not found. Install Claude Code first:\n  https://docs.anthropic.com/en/docs/claude-code\n"
+        );
+        process.exit(1);
+      }
+    } else if (currentRuntime === "codex") {
+      try {
+        execSync("codex --version", { stdio: "pipe" });
+      } catch {
+        console.error(
+          "Error: 'codex' CLI not found. Install Codex CLI first:\n  https://developers.openai.com/codex\n"
+        );
+        process.exit(1);
+      }
+    }
   }
 }
 
@@ -55,6 +75,13 @@ export function addMarketplace() {
   );
 }
 
+export function addCodexMarketplace() {
+  return run(
+    `codex plugin marketplace add ${CODEX_MARKETPLACE_PATH}`,
+    `Codex marketplace: ${CODEX_MARKETPLACE_NAME}`
+  );
+}
+
 export function installPlugin(pluginName, { scope = "project" } = {}) {
   return run(
     `claude plugin install ${pluginName}@${MARKETPLACE_NAME} --scope ${scope}`,
@@ -62,6 +89,13 @@ export function installPlugin(pluginName, { scope = "project" } = {}) {
   );
 }
 
+export function installCodexPlugin(pluginName) {
+  return run(
+    `codex plugin add ${pluginName}@${CODEX_MARKETPLACE_NAME}`,
+    pluginName
+  );
+}
+
 export function updatePlugin(pluginName, { scope = "project" } = {}) {
   return run(
     `claude plugin update "${pluginName}@${MARKETPLACE_NAME}" --scope ${scope}`,
@@ -69,18 +103,36 @@ export function updatePlugin(pluginName, { scope = "project" } = {}) {
   );
 }
 
+export function updateCodexMarketplace() {
+  return run(
+    `codex plugin marketplace add ${CODEX_MARKETPLACE_PATH}`,
+    `Codex marketplace: ${CODEX_MARKETPLACE_NAME}`
+  );
+}
+
 export function uninstallPlugin(pluginName) {
   return run(`claude plugin uninstall ${pluginName}`, `Removing ${pluginName}`);
 }
 
+export function uninstallCodexPlugin(pluginName) {
+  return run(`codex plugin remove ${pluginName}`, `Removing ${pluginName}`);
+}
+
 /**
  * Install a single package: marketplace add + plugin install.
  */
-export function installPackage(pluginName, { deps = [], scope = "project" } = {}) {
+export function installPackage(pluginName, { deps = [], scope = "project", runtime = "claude" } = {}) {
   console.log(`\nInstalling @windyroad/${pluginName.replace("wr-", "")} (${scope} scope)...\n`);
 
-  addMarketplace();
-  installPlugin(pluginName, { scope });
+  if (runtime === "claude" || runtime === "both") {
+    addMarketplace();
+    installPlugin(pluginName, { scope });
+  }
+
+  if (runtime === "codex" || runtime === "both") {
+    addCodexMarketplace();
+    installCodexPlugin(pluginName);
+  }
 
   if (deps.length > 0) {
     console.log(`\nNote: This plugin works best with:`);
@@ -90,34 +142,47 @@ export function installPackage(pluginName, { deps = [], scope = "project" } = {}
   }
 
   console.log(
-    `\nDone! Restart Claude Code to activate.\n`
+    `\nDone! Restart ${runtime === "codex" ? "Codex" : runtime === "both" ? "Claude Code and Codex" : "Claude Code"} to activate.\n`
   );
 }
 
 /**
  * Update a single package.
  */
-export function updatePackage(pluginName, { scope = "project" } = {}) {
+export function updatePackage(pluginName, { scope = "project", runtime = "claude" } = {}) {
   console.log(`\nUpdating @windyroad/${pluginName.replace("wr-", "")}...\n`);
 
-  run(
-    `claude plugin marketplace update ${MARKETPLACE_NAME}`,
-    "Updating marketplace"
-  );
-  updatePlugin(pluginName, { scope });
+  if (runtime === "claude" || runtime === "both") {
+    run(
+      `claude plugin marketplace update ${MARKETPLACE_NAME}`,
+      "Updating marketplace"
+    );
+    updatePlugin(pluginName, { scope });
+  }
 
-  console.log("\nDone! Restart Claude Code to apply updates.\n");
+  if (runtime === "codex" || runtime === "both") {
+    updateCodexMarketplace();
+    installCodexPlugin(pluginName);
+  }
+
+  console.log(`\nDone! Restart ${runtime === "codex" ? "Codex" : runtime === "both" ? "Claude Code and Codex" : "Claude Code"} to apply updates.\n`);
 }
 
 /**
  * Uninstall a single package.
  */
-export function uninstallPackage(pluginName) {
+export function uninstallPackage(pluginName, { runtime = "claude" } = {}) {
   console.log(`\nUninstalling @windyroad/${pluginName.replace("wr-", "")}...\n`);
 
-  uninstallPlugin(pluginName);
+  if (runtime === "claude" || runtime === "both") {
+    uninstallPlugin(pluginName);
+  }
+
+  if (runtime === "codex" || runtime === "both") {
+    uninstallCodexPlugin(pluginName);
+  }
 
-  console.log("\nDone. Restart Claude Code to apply changes.\n");
+  console.log(`\nDone. Restart ${runtime === "codex" ? "Codex" : runtime === "both" ? "Claude Code and Codex" : "Claude Code"} to apply changes.\n`);
 }
 
 /**
@@ -131,6 +196,7 @@ export function parseStandardArgs(argv) {
     update: args.includes("--update"),
     dryRun: args.includes("--dry-run"),
     scope: "project",
+    runtime: "claude",
   };
   const scopeIdx = args.indexOf("--scope");
   if (scopeIdx !== -1 && args[scopeIdx + 1]) {
@@ -142,5 +208,15 @@ export function parseStandardArgs(argv) {
       process.exit(1);
     }
   }
+  const runtimeIdx = args.indexOf("--runtime");
+  if (runtimeIdx !== -1 && args[runtimeIdx + 1]) {
+    const val = args[runtimeIdx + 1];
+    if (["claude", "codex", "both"].includes(val)) {
+      flags.runtime = val;
+    } else {
+      console.error("--runtime requires: claude, codex, or both");
+      process.exit(1);
+    }
+  }
   return flags;
 }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@windyroad/architect",
-  "version": "0.17.3-preview.755",
+  "version": "0.17.3-preview.766",
   "description": "Architecture decision enforcement for AI coding agents",
   "bin": {
     "windyroad-architect": "./bin/install.mjs"
@@ -25,6 +25,7 @@
     "skills/",
     "scripts/",
     ".claude-plugin/",
+    ".codex-plugin/",
     "lib/",
     "!skills/*/eval/"
   ]

package/skills/create-adr/test/create-adr-codex-eval-contract.bats

@@ -0,0 +1,30 @@
+#!/usr/bin/env bats
+
+setup() {
+  REPO_ROOT="$(cd "$(dirname "$BATS_TEST_FILENAME")/../../../../.." && pwd)"
+  EVAL_DIR="$REPO_ROOT/packages/architect/skills/create-adr/eval"
+}
+
+@test "create-adr Codex promptfoo config sits beside Claude config" {
+  [ -f "$EVAL_DIR/promptfooconfig.yaml" ]
+  [ -f "$EVAL_DIR/promptfooconfig.codex.yaml" ]
+  run grep -n "run-codex-skill-eval.sh" "$EVAL_DIR/promptfooconfig.codex.yaml"
+  [ "$status" -eq 0 ]
+  run grep -n "grade-codex-rubric.sh" "$EVAL_DIR/promptfooconfig.codex.yaml"
+  [ "$status" -eq 0 ]
+}
+
+@test "create-adr Codex eval runner uses codex exec and plugin marketplace" {
+  run grep -n "codex plugin marketplace add" "$EVAL_DIR/run-codex-skill-eval.sh"
+  [ "$status" -eq 0 ]
+  run grep -n "codex plugin add wr-architect@windyroad-local" "$EVAL_DIR/run-codex-skill-eval.sh"
+  [ "$status" -eq 0 ]
+  run grep -n "codex exec" "$EVAL_DIR/run-codex-skill-eval.sh"
+  [ "$status" -eq 0 ]
+  run grep -n -- "--ephemeral" "$EVAL_DIR/run-codex-skill-eval.sh"
+  [ "$status" -eq 0 ]
+  run grep -n -- 'approval_policy="never"' "$EVAL_DIR/run-codex-skill-eval.sh"
+  [ "$status" -eq 0 ]
+  run grep -n -- "--sandbox read-only" "$EVAL_DIR/run-codex-skill-eval.sh"
+  [ "$status" -eq 0 ]
+}

package/skills/review-design/SKILL.md

@@ -69,16 +69,27 @@ Build a self-contained prompt for the architect subagent that includes:
 
 The architect's verdict taxonomy includes **[Unratified Dependency]** (ADR-074 surface 3): if the plan/change explicitly cites or implements an ADR that lacks `human-oversight: confirmed` (unratified, non-superseded), the architect flags ISSUES FOUND with a "ratify via /wr-architect:review-decisions first" action. This applies to plan review exactly as to edit review — a plan built on an unratified decision should not proceed until that decision's substance is ratified. No extra prompt wiring is needed (the agent owns the check); this note records that the surface-3 check is in-scope for plan review.
 
-### 5. Delegate to wr-architect:agent
+### 5. Delegate to the architect agent
 
-Invoke the architect subagent via the `Skill` tool:
+Use the runtime-native architect agent path with the constructed prompt.
+
+Claude Code invocation:
 
 ```
 subagent_type: wr-architect:agent
 prompt: <constructed review prompt from step 4>
 ```
 
-Wait for the subagent to complete.
+Codex invocation:
+
+```
+agent: wr-architect:agent
+prompt: <constructed review prompt from step 4>
+```
+
+Codex auto-discovers the agent from the installed plugin's `agents/agent.md` and registers it as `wr-architect:agent` (plugin name + agent frontmatter name). Adopter projects do not need to ship any `.codex/agents/*.toml` themselves — `codex plugin add wr-architect@windyroad` is enough. The repo-local `.codex/agents/wr-architect.toml`, kept in sync from `packages/architect/agents/agent.md` by `scripts/sync-codex-agents.mjs`, only supports in-repo dogfooding before the plugin is installed into `CODEX_HOME`; it registers a short-name alias `wr-architect` for source-repo developers. If a non-interactive `codex exec` validation harness cannot spawn the custom agent directly, perform the review inline using the same generated agent instructions and name that fallback in the output.
+
+Wait for the agent to complete.
 
 ### 6. Present results
 
@@ -87,7 +98,7 @@ Present the full compliance report to the user. The architect subagent will repo
 - FLAGGED: specific violations or compliance questions with ADR references
 - NEW ADR NEEDED: decisions that should be recorded before proceeding
 
-If violations are flagged, use `AskUserQuestion` to ask how the user wants to proceed:
+If violations are flagged, use the runtime's structured question mechanism to ask how the user wants to proceed:
 - (a) Address the violations before continuing
 - (b) Proceed with a documented exception
 - (c) Draft a new or amended ADR to legitimise the approach

package/skills/review-design/test/review-design-codex-contract.bats

@@ -0,0 +1,22 @@
+#!/usr/bin/env bats
+
+setup() {
+  REPO_ROOT="$(cd "$(dirname "$BATS_TEST_FILENAME")/../../../../.." && pwd)"
+  SKILL="$REPO_ROOT/packages/architect/skills/review-design/SKILL.md"
+}
+
+@test "review-design documents Claude and Codex architect invocation paths" {
+  run grep -n "Claude Code invocation" "$SKILL"
+  [ "$status" -eq 0 ]
+  run grep -n "subagent_type: wr-architect:agent" "$SKILL"
+  [ "$status" -eq 0 ]
+  run grep -n "Codex invocation" "$SKILL"
+  [ "$status" -eq 0 ]
+  run grep -nE "^agent: wr-architect:agent$" "$SKILL"
+  [ "$status" -eq 0 ]
+}
+
+@test "review-design names generated Codex agent config" {
+  run grep -n ".codex/agents/wr-architect.toml" "$SKILL"
+  [ "$status" -eq 0 ]
+}