@raishin/vanguard-frontier-agentic 2.7.1 → 2.9.0

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@raishin/vanguard-frontier-agentic",
-  "version": "2.7.1",
+  "version": "2.9.0",
   "description": "Cloud and zero-trust agentic workflow marketplace for skills, agents, rules, MCP references, and compliance-aware architecture.",
   "license": "Apache-2.0",
   "repository": {
@@ -39,6 +39,7 @@
     "agents:export": "node scripts/export-marketplace-agents.mjs",
     "install:codex-home": "node scripts/install-codex-home.mjs",
     "manifest:write": "python3 tests/validate-skill-manifest.py --write",
+    "docs-data:write": "node scripts/generate-docs-data.mjs",
     "manifest:check": "python3 tests/validate-skill-manifest.py",
     "validate:aws": "python3 tests/validate-aws-skill-quality.py && python3 tests/validate-aws-progressive-disclosure.py",
     "validate:catalog": "python3 tests/validate-catalog.py",

package/plugins/vanguard-frontier-agentic/.codex-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "vanguard-frontier-agentic",
-  "version": "2.7.1",
+  "version": "2.9.0",
   "description": "Curated marketplace for cloud and zero-trust AI workflows. 331 agents, 286 skills, and rules across AWS, Azure, OCI, GCP, Alibaba Cloud, Huawei Cloud, Kubernetes, and Terraform.",
   "author": {
     "name": "Raishin",

package/powers/README.md

@@ -1,6 +1,6 @@
 # `powers/` — Kiro Powers
 
-This directory holds **32 Kiro Powers** for `vanguard-frontier-agentic`, one
+This directory holds **34 Kiro Powers** for `vanguard-frontier-agentic`, one
 per cloud/platform/IaC provider. Each Power is a directory containing a
 `POWER.md` file with strict-5 frontmatter and steering content.
 
@@ -8,6 +8,7 @@ per cloud/platform/IaC provider. Each Power is a directory containing a
 
 ```
 powers/
+├── vanguard-accounting/POWER.md
 ├── vanguard-alibaba/POWER.md
 ├── vanguard-argocd/POWER.md
 ├── vanguard-aws/POWER.md
@@ -18,6 +19,7 @@ powers/
 ├── vanguard-contabo/POWER.md
 ├── vanguard-dotnet/POWER.md
 ├── vanguard-falco/POWER.md
+├── vanguard-finance/POWER.md
 ├── vanguard-fluxcd/POWER.md
 ├── vanguard-gcp/POWER.md
 ├── vanguard-generic/POWER.md
@@ -76,7 +78,7 @@ cd vanguard-frontier-agentic
 ## How to update
 
 ```bash
-# Regenerate the 32 Powers from catalog/agents.json + per-provider config:
+# Regenerate the 34 Powers from catalog/agents.json + per-provider config:
 npm run kiro-powers:write
 
 # Then verify everything is in sync:

package/powers/vanguard-accounting/POWER.md

@@ -0,0 +1,40 @@
+---
+name: "vanguard-accounting"
+displayName: "Vanguard Frontier — Accounting"
+description: "Curated Accounting agents for accounting revenue recognition advisor, accounting close cycle advisor, accounting tax provision advisor, accounting lease accounting advisor. Routes via accounting-maestro-agent to specialist agents based on task scope. Static review only; no live mutations."
+keywords: ["accounting", "static-review", "configuration-audit", "best-practices"]
+author: "Raishin"
+---
+# Vanguard Frontier — Accounting
+
+Curated Accounting agents for accounting revenue recognition advisor, accounting close cycle advisor, accounting tax provision advisor, accounting lease accounting advisor. Routes via accounting-maestro-agent to specialist agents based on task scope. Static review only; no live mutations.
+
+## When to engage this Power
+
+Activate when the task references Accounting services, resources, or operations. Do not activate on unrelated requests — narrow keyword matching is required to avoid false activations (Kiro Powers convention).
+
+## Routing pattern
+
+- **`accounting-maestro-agent`** — classifies and routes the task to the right specialist
+
+Use the maestro as the entry point: classify the task, then dispatch to one specialist or a parallel team of specialists. Never have the maestro itself execute a live mutation.
+
+## Live-guard agents (gate_mode only)
+
+- *(none — this provider has no live-mutation guards in the catalog)*
+
+Live-guard agents enforce approval, target confirmation, evidence capture, and rollback plans before executing a mutation. They are never auto-dispatched — the maestro must place them in `live-guard-gate` or `runtime-evidence-gate` mode.
+
+## Invariants
+
+- Route all tasks through accounting-maestro-agent for proper classification and dispatch.
+- Static review only -- agents analyze configuration and provide findings without mutating live systems.
+
+## Where the agents live
+
+Agent specs and adapters are part of the [Vanguard Frontier Agentic](https://github.com/Raishin/vanguard-frontier-agentic) marketplace. For this provider, see `agents/accounting/` in that repository. All 14 agents in this provider ship a Kiro adapter (`harnesses/kiro-ide.agent.md`, `kiro-cli.agent.json`).
+
+## Companion install paths
+
+- **Claude Code:** `/plugin marketplace add Raishin/vanguard-frontier-agentic` then `/plugin install vanguard-frontier-agentic@vanguard-frontier-agentic`
+- **Codex / Copilot / Cursor / Gemini CLI / Kiro (file export):** `npx vfa-export-agents --platform <harness> --provider accounting --repo .`

package/powers/vanguard-finance/POWER.md

@@ -0,0 +1,40 @@
+---
+name: "vanguard-finance"
+displayName: "Vanguard Frontier — Finance"
+description: "Curated Finance agents for finance variance analysis advisor, finance treasury liquidity advisor, finance capital allocation advisor, finance transfer pricing pillar two advisor. Routes via finance-maestro-agent to specialist agents based on task scope. Static review only; no live mutations."
+keywords: ["finance", "static-review", "configuration-audit", "best-practices"]
+author: "Raishin"
+---
+# Vanguard Frontier — Finance
+
+Curated Finance agents for finance variance analysis advisor, finance treasury liquidity advisor, finance capital allocation advisor, finance transfer pricing pillar two advisor. Routes via finance-maestro-agent to specialist agents based on task scope. Static review only; no live mutations.
+
+## When to engage this Power
+
+Activate when the task references Finance services, resources, or operations. Do not activate on unrelated requests — narrow keyword matching is required to avoid false activations (Kiro Powers convention).
+
+## Routing pattern
+
+- **`finance-maestro-agent`** — classifies and routes the task to the right specialist
+
+Use the maestro as the entry point: classify the task, then dispatch to one specialist or a parallel team of specialists. Never have the maestro itself execute a live mutation.
+
+## Live-guard agents (gate_mode only)
+
+- *(none — this provider has no live-mutation guards in the catalog)*
+
+Live-guard agents enforce approval, target confirmation, evidence capture, and rollback plans before executing a mutation. They are never auto-dispatched — the maestro must place them in `live-guard-gate` or `runtime-evidence-gate` mode.
+
+## Invariants
+
+- Route all tasks through finance-maestro-agent for proper classification and dispatch.
+- Static review only -- agents analyze configuration and provide findings without mutating live systems.
+
+## Where the agents live
+
+Agent specs and adapters are part of the [Vanguard Frontier Agentic](https://github.com/Raishin/vanguard-frontier-agentic) marketplace. For this provider, see `agents/finance/` in that repository. All 8 agents in this provider ship a Kiro adapter (`harnesses/kiro-ide.agent.md`, `kiro-cli.agent.json`).
+
+## Companion install paths
+
+- **Claude Code:** `/plugin marketplace add Raishin/vanguard-frontier-agentic` then `/plugin install vanguard-frontier-agentic@vanguard-frontier-agentic`
+- **Codex / Copilot / Cursor / Gemini CLI / Kiro (file export):** `npx vfa-export-agents --platform <harness> --provider finance --repo .`

package/schemas/agent.schema.json

@@ -53,7 +53,9 @@
         "dotnet",
         "hr",
         "legal",
-        "salesforce"
+        "salesforce",
+        "accounting",
+        "finance"
       ]
     },
     "harnesses": {
@@ -111,18 +113,30 @@
       }
     },
     "execution_tier": {
-      "enum": ["static-review", "read-only-runtime", "mutating-runtime"],
+      "enum": [
+        "static-review",
+        "read-only-runtime",
+        "mutating-runtime"
+      ],
       "description": "Trust posture: static-review (no Bash), read-only-runtime (allowlisted Bash, no mutation), mutating-runtime (changes external state)."
     },
     "lifecycle": {
-      "enum": ["experimental", "beta", "stable", "deprecated"],
+      "enum": [
+        "experimental",
+        "beta",
+        "stable",
+        "deprecated"
+      ],
       "description": "Stability stage. Defaults to 'stable' when absent."
     },
     "harness_variants": {
       "type": "object",
       "description": "Per-harness adapter file paths. Keys are harness ids (claude-code, cursor, codex, copilot, gemini, kiro-cli, kiro-ide).",
-      "additionalProperties": { "type": "string", "minLength": 1 }
+      "additionalProperties": {
+        "type": "string",
+        "minLength": 1
+      }
     }
   },
   "additionalProperties": true
-}
+}

package/schemas/skill.frontmatter.schema.json

@@ -18,7 +18,7 @@
     },
     "description": {
       "type": "string",
-      "description": "Human-readable purpose statement. 50–1500 characters.",
+      "description": "Human-readable purpose statement. 50\u20131500 characters.",
       "minLength": 50,
       "maxLength": 1500
     },
@@ -85,13 +85,19 @@
             "cost-management",
             "operational",
             "generation",
-            "devsecops"
+            "devsecops",
+            "finance"
           ]
         },
         "lifecycle": {
           "type": "string",
           "description": "Stability stage. Optional; defaults to 'stable' when absent.",
-          "enum": ["experimental", "beta", "stable", "deprecated"]
+          "enum": [
+            "experimental",
+            "beta",
+            "stable",
+            "deprecated"
+          ]
         }
       },
       "additionalProperties": true
@@ -107,18 +113,29 @@
       "description": "Optional fields for live-execution skills (e.g. read-only-runtime gates). Declared at the metadata block level. See docs/live-agents/ for the full contract.",
       "properties": {
         "execution_tier": {
-          "enum": ["static-review", "read-only-runtime", "sandbox-mutating", "mutating-runtime"],
-          "description": "Trust posture: static-review (no execution), read-only-runtime (allowlisted Bash, no mutation), sandbox-mutating (dry-run or sandbox-only changes, never production), mutating-runtime (changes external state — gated by additional human-in-the-loop controls)."
+          "enum": [
+            "static-review",
+            "read-only-runtime",
+            "sandbox-mutating",
+            "mutating-runtime"
+          ],
+          "description": "Trust posture: static-review (no execution), read-only-runtime (allowlisted Bash, no mutation), sandbox-mutating (dry-run or sandbox-only changes, never production), mutating-runtime (changes external state \u2014 gated by additional human-in-the-loop controls)."
         },
         "mcp_servers": {
           "type": "array",
-          "items": { "type": "string", "minLength": 1 },
+          "items": {
+            "type": "string",
+            "minLength": 1
+          },
           "description": "Optional. List of MCP server identifiers this skill targets. Empty array means no MCP integration (CLI-only or static-review)."
         },
         "oauth_scopes": {
           "type": "array",
-          "items": { "type": "string", "minLength": 1 },
-          "description": "Optional. Explicit list of OAuth scopes required for live execution. Empty array means no OAuth scopes needed (static-review or local-only operations). For Salesforce: minimum scopes only — never request 'full' or 'web'."
+          "items": {
+            "type": "string",
+            "minLength": 1
+          },
+          "description": "Optional. Explicit list of OAuth scopes required for live execution. Empty array means no OAuth scopes needed (static-review or local-only operations). For Salesforce: minimum scopes only \u2014 never request 'full' or 'web'."
         },
         "run_as_permissions": {
           "type": "object",
@@ -126,12 +143,18 @@
           "properties": {
             "required": {
               "type": "array",
-              "items": { "type": "string", "minLength": 1 },
+              "items": {
+                "type": "string",
+                "minLength": 1
+              },
               "description": "Minimum permission set or profile permissions required on the Run As account."
             },
             "denied": {
               "type": "array",
-              "items": { "type": "string", "minLength": 1 },
+              "items": {
+                "type": "string",
+                "minLength": 1
+              },
               "description": "Permissions that MUST be explicitly denied on the Run As account. For Salesforce T1: ModifyAllData, ViewAllData, ViewEncryptedData, ModifyMetadata are typical denies."
             }
           },
@@ -139,19 +162,34 @@
         },
         "required_egress": {
           "type": "array",
-          "items": { "type": "string", "minLength": 1 },
+          "items": {
+            "type": "string",
+            "minLength": 1
+          },
           "description": "Named hostnames the skill may contact in runtime mode. Empty array means no egress."
         },
         "requires_credentials": {
           "type": "array",
-          "items": { "type": "string", "minLength": 1 },
+          "items": {
+            "type": "string",
+            "minLength": 1
+          },
           "description": "Environment variable names the skill may read. Values are never collected, never echoed."
         },
         "output_attestation": {
           "type": "object",
           "properties": {
-            "schema": { "type": "string", "minLength": 1 },
-            "signed_with": { "enum": ["cosign-blob", "in-toto", "none"] }
+            "schema": {
+              "type": "string",
+              "minLength": 1
+            },
+            "signed_with": {
+              "enum": [
+                "cosign-blob",
+                "in-toto",
+                "none"
+              ]
+            }
           },
           "additionalProperties": false
         },
@@ -163,4 +201,4 @@
     }
   },
   "additionalProperties": true
-}
+}

package/scripts/generate-docs-data.mjs

@@ -0,0 +1,116 @@
+#!/usr/bin/env node
+/**
+ * Generate docs/_data/catalog.yml from the live catalog.
+ *
+ * Single source of truth for all counts displayed on the Jekyll documentation site.
+ * Run this after any catalog change to keep docs in sync.
+ *
+ * Usage: node scripts/generate-docs-data.mjs
+ */
+
+import { readFileSync, writeFileSync, readdirSync, mkdirSync } from 'fs';
+import { join, dirname } from 'path';
+import { fileURLToPath } from 'url';
+
+const __dirname = dirname(fileURLToPath(import.meta.url));
+const root = join(__dirname, '..');
+
+// Load catalog data
+const agents = JSON.parse(readFileSync(join(root, 'catalog/agents.json'), 'utf8'));
+const skills = JSON.parse(readFileSync(join(root, 'catalog/skills.json'), 'utf8'));
+const mcpRefs = JSON.parse(readFileSync(join(root, 'catalog/mcp-references.json'), 'utf8'));
+const rules = JSON.parse(readFileSync(join(root, 'catalog/rules.json'), 'utf8'));
+const roles = JSON.parse(readFileSync(join(root, 'catalog/install-roles.json'), 'utf8'));
+const pkg = JSON.parse(readFileSync(join(root, 'package.json'), 'utf8'));
+
+// Compute counts
+const providers = [...new Set(agents.map(a => a.provider))].sort();
+const validateScripts = Object.keys(pkg.scripts || {}).filter(k => k.startsWith('validate:'));
+const agentDirs = readdirSync(join(root, 'agents'));
+const skillDirs = readdirSync(join(root, 'skills'));
+const roleIds = Object.keys(roles.roles).sort();
+
+// Count maestro routing scenarios from the test output or fixture
+let maestroScenarios = 357; // default
+try {
+  const maestroFixture = readFileSync(join(root, 'tests/fixtures/maestro-routing-scenarios.json'), 'utf8');
+  const scenarios = JSON.parse(maestroFixture);
+  maestroScenarios = Array.isArray(scenarios) ? scenarios.length : 357;
+} catch {
+  // Use default if fixture doesn't exist
+}
+
+// Generate YAML
+// Compute provider taxonomy (categorize providers for documentation)
+const providerCounts = {};
+for (const agent of agents) {
+  providerCounts[agent.provider] = (providerCounts[agent.provider] || 0) + 1;
+}
+
+const taxonomy = [
+  { category: 'Cloud Hyperscalers', providers: ['aws', 'azure', 'gcp', 'oci', 'alibaba', 'huawei'] },
+  { category: 'European Cloud', providers: ['ovhcloud', 'ionos', 'scaleway', 'hetzner', 'contabo'] },
+  { category: 'Container & Orchestration', providers: ['kubernetes', 'argocd', 'fluxcd', 'istio', 'cilium'] },
+  { category: 'Security & Supply Chain', providers: ['falco', 'kyverno', 'sigstore', 'cert-manager'] },
+  { category: 'Observability', providers: ['opentelemetry', 'prometheus'] },
+  { category: 'Infrastructure as Code', providers: ['terraform'] },
+  { category: 'AI & Compute', providers: ['nvidia'] },
+  { category: 'Developer Platforms', providers: ['backstage', 'dotnet', 'generic', 'multi-cloud'] },
+  { category: 'Business Functions', providers: ['salesforce', 'legal', 'hr', 'marketing'] },
+];
+
+const taxonomyYaml = taxonomy.map(t => {
+  const entries = t.providers
+    .filter(p => providerCounts[p] > 0)
+    .map(p => `      - name: "${p}"\n        agents: ${providerCounts[p]}`)
+    .join('\n');
+  return `  - category: "${t.category}"\n    providers:\n${entries}`;
+}).join('\n');
+
+const yaml = `# Auto-generated by scripts/generate-docs-data.mjs
+# Do NOT edit manually — run: node scripts/generate-docs-data.mjs
+# This file is the single source of truth for all catalog counts in Jekyll docs.
+
+agents: ${agents.length}
+skills: ${skills.length}
+providers: ${providers.length}
+mcp_references: ${mcpRefs.length}
+rules: ${rules.length}
+install_roles: ${roleIds.length}
+validation_gates: ${validateScripts.length}
+agent_directories: ${agentDirs.length}
+skill_directories: ${skillDirs.length}
+maestro_scenarios: ${maestroScenarios}
+version: "${pkg.version}"
+
+# Provider list (flat)
+provider_list:
+${providers.map(p => `  - "${p}"`).join('\n')}
+
+# Provider taxonomy (grouped by category with agent counts)
+provider_taxonomy:
+${taxonomyYaml}
+
+# Role list
+role_list:
+${roleIds.map(r => `  - "${r}"`).join('\n')}
+
+# Harness list
+harness_list:
+  - "claude-code"
+  - "codex"
+  - "copilot"
+  - "cursor"
+  - "gemini"
+  - "kiro"
+`;
+
+// Write output
+const outDir = join(root, 'docs/_data');
+mkdirSync(outDir, { recursive: true });
+const outPath = join(outDir, 'catalog.yml');
+writeFileSync(outPath, yaml);
+
+console.log(`OK: wrote ${outPath}`);
+console.log(`  agents=${agents.length} skills=${skills.length} providers=${providers.length} roles=${roleIds.length} gates=${validateScripts.length}`);
+

package/scripts/generate-plugin-manifest.mjs

@@ -1,6 +1,7 @@
 #!/usr/bin/env node
 /**
- * Generate .claude-plugin/plugin.json from catalog/agents.json.
+ * Generate .claude-plugin/plugin.json and .claude-plugin/marketplace.json
+ * from catalog/agents.json and package.json.
  *
  * The Claude Code plugin spec lets a manifest declare its agents via an
  * explicit array of file paths (see code.claude.com/docs/en/plugins-reference,
@@ -12,6 +13,9 @@
  * design and the npm package layout), this script enumerates every
  * claude-code adapter file and writes it into plugin.json's `agents` array.
  *
+ * marketplace.json version and agent count are also auto-computed so both
+ * files stay in sync with package.json as the single source of truth.
+ *
  * Output is sorted and deterministic so the manifest is reproducible across
  * runs and reviewable in PR diffs.
  *
@@ -28,6 +32,7 @@ const repoRoot = join(dirname(fileURLToPath(import.meta.url)), "..");
 const catalogPath = join(repoRoot, "catalog", "agents.json");
 const skillManifestPath = join(repoRoot, "catalog", "skill-manifest.json");
 const manifestPath = join(repoRoot, ".claude-plugin", "plugin.json");
+const marketplacePath = join(repoRoot, ".claude-plugin", "marketplace.json");
 const pkgPath = join(repoRoot, "package.json");
 
 const check = process.argv.includes("--check");
@@ -66,8 +71,8 @@ function manifestPathForAdapter(entry, adapter) {
   return `./${rel.split(sep).join("/")}`;
 }
 
-const agentEntries = catalog
-  .filter((entry) => entry.type === "agent")
+const allAgents = catalog.filter((entry) => entry.type === "agent");
+const agentEntries = allAgents
   .filter((entry) => Array.isArray(entry.harnesses) && entry.harnesses.includes("claude-code"))
   .map((entry) => {
     const adapter = entry.harness_variants?.["claude-code"]
@@ -95,6 +100,10 @@ const skillCount = existsSync(skillManifestPath)
   ? JSON.parse(readFileSync(skillManifestPath, "utf8")).length ?? 0
   : 0;
 
+// Derive a concise agent count and provider list for the marketplace description.
+const totalAgents = agentEntries.length;
+const providers = [...new Set(allAgents.map((a) => a.provider).filter(Boolean))].sort();
+
 const manifest = {
   name: "vanguard-frontier-agentic",
   version: pkg.version,
@@ -126,26 +135,79 @@ const manifest = {
   agents: agentEntries,
 };
 
-const next = JSON.stringify(manifest, null, 2) + "\n";
+// marketplace.json — version and agent count auto-computed from package.json
+// and catalog/agents.json so they never diverge from the installed artifact.
+const marketplace = {
+  name: "vanguard-frontier-agentic",
+  owner: {
+    name: "Raishin",
+    url: "https://github.com/Raishin",
+  },
+  metadata: {
+    description: pkg.description,
+    version: pkg.version,
+  },
+  plugins: [
+    {
+      name: "vanguard-frontier-agentic",
+      source: "./",
+      description: `All ${totalAgents} cloud, security, compliance, platform, accounting, and finance agents in one install. Includes maestros, advisory reviewers, and live-mutation guards across ${providers.length} providers.`,
+      category: "cloud",
+      tags: [
+        "agents",
+        "cloud",
+        "kubernetes",
+        "terraform",
+        "zero-trust",
+        "compliance",
+        "live-guards",
+      ],
+      strict: true,
+    },
+  ],
+};
+
+const nextManifest = JSON.stringify(manifest, null, 2) + "\n";
+const nextMarketplace = JSON.stringify(marketplace, null, 2) + "\n";
 
 if (check) {
+  let ok = true;
   if (!existsSync(manifestPath)) {
     console.error(`ERROR: ${manifestPath} is missing; run npm run plugin-manifest:write`);
-    process.exit(1);
+    ok = false;
+  } else {
+    const current = readFileSync(manifestPath, "utf8");
+    if (current !== nextManifest) {
+      console.error(
+        `ERROR: ${relative(repoRoot, manifestPath)} is stale (${agentEntries.length} agents in catalog, manifest is out of sync); run npm run plugin-manifest:write`,
+      );
+      ok = false;
+    }
   }
-  const current = readFileSync(manifestPath, "utf8");
-  if (current !== next) {
-    console.error(
-      `ERROR: ${relative(repoRoot, manifestPath)} is stale (${agentEntries.length} agents in catalog, manifest is out of sync); run npm run plugin-manifest:write`,
-    );
-    process.exit(1);
+  if (!existsSync(marketplacePath)) {
+    console.error(`ERROR: ${marketplacePath} is missing; run npm run plugin-manifest:write`);
+    ok = false;
+  } else {
+    const currentMarketplace = readFileSync(marketplacePath, "utf8");
+    if (currentMarketplace !== nextMarketplace) {
+      console.error(
+        `ERROR: ${relative(repoRoot, marketplacePath)} is stale (version or agent count changed); run npm run plugin-manifest:write`,
+      );
+      ok = false;
+    }
   }
+  if (!ok) process.exit(1);
   console.log(
     `OK: plugin manifest is in sync (${agentEntries.length} agents, ${skillCount} skills tracked in catalog/skill-manifest.json)`,
   );
 } else {
-  writeFileSync(manifestPath, next);
+  writeFileSync(manifestPath, nextManifest);
+  writeFileSync(marketplacePath, nextMarketplace);
   console.log(
     `OK: wrote ${relative(repoRoot, manifestPath)} (${agentEntries.length} agents)`,
   );
+  console.log(
+    `OK: wrote ${relative(repoRoot, marketplacePath)} (version=${pkg.version}, agents=${totalAgents}, providers=${providers.length})`,
+  );
 }
+

package/skills/accounting/README.md

@@ -0,0 +1,8 @@
+# Accounting Skills
+
+Reference skills for corporate accounting advisory workflows grounded in GAAP (ASC) and IFRS standards.
+
+| Skill | Purpose |
+|---|---|
+| [accounting-maestro](./accounting-maestro/) | Routing logic and safety protocol for accounting domain |
+| [revenue-recognition-advisor](./revenue-recognition-advisor/) | ASC 606 / IFRS 15 five-step model application framework |

package/skills/accounting/accounting-maestro/README.md

@@ -0,0 +1,14 @@
+# Accounting Maestro Skill
+
+Routing logic and safety protocol for the accounting domain. Used by the `accounting-maestro-agent` to classify and dispatch accounting tasks to the correct specialist.
+
+## What it does
+
+- Maps accounting task types to catalog agent IDs
+- Enforces routing rules: single vs. parallel dispatch, hard ceiling, live-guard gate
+- Provides the safety checklist for dispatch validation
+- Establishes the response shape for route/reason/mode headers
+
+## Domain coverage
+
+Revenue recognition → close cycle → audit evidence → consolidation (as catalog expands)