chore(release): v5.13.5 - axiom v0.6.2 pin + numeric sync

- Pin axiom marketplace entry to v0.6.2 / e3f3fab. v0.6.2 ships the
  SessionStart plugin hook, the read-only `before-any --detect-only`
  mode, an OpenCode `session.created` plugin scaffold, and the
  `BeforeAnyError` library refactor.
- Sync agent / skill / test / rule counts across README.md, AGENTS.md,
  docs/reference/AGENTS.md, .codex-plugin/plugin.json, site/content.json,
  site/ux-spec.md, and site/index.html. The drifted numbers were:
  - File-based agents: 39 -> 40 (total 49 -> 50). Generator
    STATIC_PLUGIN_AGENT_COUNTS map updated for repo-intel (1 -> 3) and
    agnix (1 -> 0) to match the actual filesystem.
  - Tests: 3,513 -> 3,518.
  - agnix: 399 -> 423 rules, 126 -> 129 auto-fixable.
- Update /axiom section in README to document the SessionStart hook,
  --detect-only flag, Codex plugin-hooks opt-in
  (`[features].plugin_hooks = true`), and the OpenCode scaffold
  limitation pending anomalyco/opencode#5409.
- Update marketplace-standalone-contract.test.js to pin axiom at
  v0.6.2 / e3f3fab.

Tests: 87 suites, 3,518 passing, 39 skipped, 0 failing.

Author: avifenesh

.claude-plugin/marketplace.json

@@ -1,7 +1,7 @@
 {
   "name": "agentsys",
   "description": "24 specialized plugins for AI workflow automation - task orchestration, PR workflow, slop detection, code review, drift detection, enhancement analysis, documentation sync, unified static analysis, durable memory, negative behavior memory, skill and system prompt curation, perf investigations, topic research, agent config linting, cross-tool AI consultation, structured AI debate, workflow pattern learning, codebase onboarding, contributor guidance, and Zig language support",
-  "version": "5.13.4",
+  "version": "5.13.5",
   "owner": {
     "name": "Avi Fenesh",
     "url": "https://github.com/avifenesh"
@@ -203,11 +203,11 @@
       "source": {
         "source": "url",
         "url": "https://github.com/agent-sh/axiom.git",
-        "ref": "v0.6.1",
-        "commit": "ba77271a4ff7084d2d567cf3173204390be80f3b"
+        "ref": "v0.6.2",
+        "commit": "e3f3fabfcb19c140a38cbd5c5129ac69ca50b359"
       },
       "description": "Personal agent-native knowledge base: load thin context, query durable memories, create project scaffolds, and propose approved records",
-      "version": "0.6.1",
+      "version": "0.6.2",
       "category": "productivity",
       "homepage": "https://github.com/agent-sh/axiom"
     },

.claude-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "agentsys",
-  "version": "5.13.4",
+  "version": "5.13.5",
   "description": "Professional-grade slash commands for Claude Code with cross-platform support",
   "keywords": [
     "workflow",

.codex-plugin/plugin.json

@@ -4,7 +4,7 @@
   "skills": "./adapters/codex/skills",
   "interface": {
     "displayName": "agentsys",
-    "shortDescription": "AI agent orchestration with 45 skills and 49 agents",
+    "shortDescription": "AI agent orchestration with 45 skills and 50 agents",
     "longDescription": "Professional-grade slash commands for AI-powered development workflows. Includes /next-task (task discovery to production), /axiom (durable agent memory), /banthis (negative behavior memory), /skill-curator and /system-prompt-curator (agent configuration curation), /ship (commit to deploy), /audit-project (multi-agent code review), /deslop (AI slop cleanup), /perf (performance investigation), /enhance (config analysis), /consult (cross-tool AI consultation), and more.",
     "developerName": "Avi Fenesh",
     "category": "developer-tools",

AGENTS.md

@@ -76,7 +76,7 @@
 <!-- GEN:START:claude-architecture -->
 ```
 lib/          → Shared library (vendored to plugins)
-plugins/      → 24 plugins, 49 agents (39 file-based + 10 role-based), 45 skills
+plugins/      → 24 plugins, 50 agents (40 file-based + 10 role-based), 45 skills
 adapters/     → Platform adapters (opencode-plugin/, opencode/, codex/)
 checklists/   → Action checklists (9 files)
 bin/cli.js    → npm CLI installer
@@ -93,12 +93,12 @@ bin/cli.js    → npm CLI installer
 | drift-detect | 1 | 1 | Plan drift detection |
 | enhance | 8 | 9 | Code quality analyzers |
 | sync-docs | 1 | 1 | Documentation sync |
-| repo-intel | 1 | 1 | Unified static analysis |
+| repo-intel | 3 | 1 | Unified static analysis |
 | axiom | 0 | 1 | Durable agent-native memory |
 | banthis | 0 | 1 | Durable negative behavior memory |
 | perf | 6 | 8 | Performance investigation |
 | learn | 1 | 1 | Topic research and learning guides |
-| agnix | 1 | 1 | Agent config linting |
+| agnix | 0 | 1 | Agent config linting |
 | consult | 1 | 1 | Cross-tool AI consultation |
 | debate | 1 | 1 | Multi-perspective debate analysis |
 | web-ctl | 1 | 2 | Browser automation for AI agents |
@@ -176,7 +176,7 @@ agentsys                # Run installer
 <agents>
 ## Agents
 
-49 agents across 24 plugins (18 have agents; gate-and-ship is commands-only; axiom, banthis, skill-curator, and system-prompt-curator are skill/command-only; zig-lsp is config-only with no commands or agents). Key agents by model:
+50 agents across 24 plugins (18 have agents; gate-and-ship is commands-only; axiom, banthis, skill-curator, and system-prompt-curator are skill/command-only; zig-lsp is config-only with no commands or agents). Key agents by model:
 
 | Model | Agents | Use Case |
 |-------|--------|----------|

CHANGELOG.md

@@ -9,6 +9,18 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [5.13.5] - 2026-05-18
+
+### Changed
+- Updated `axiom` marketplace pin to `v0.6.2`, which ships the `SessionStart` plugin hook (`hooks/hooks.json` + `hooks/session-start.mjs`), the read-only `axiom before-any --detect-only` mode, an OpenCode `session.created` plugin scaffold, and the `BeforeAnyError` library refactor.
+
+### Fixed
+- Synced agent / skill / test / rule counts across `README.md`, `AGENTS.md`, `docs/reference/AGENTS.md`, `.codex-plugin/plugin.json`, `site/content.json`, `site/ux-spec.md`, and `site/index.html`. File-based agent count corrected `39 → 40` (total `49 → 50`), test count refreshed `3,513 → 3,518`, agnix references refreshed `399 → 423` rules and `126 → 129` auto-fixable. Updated `STATIC_PLUGIN_AGENT_COUNTS` in `scripts/generate-docs.js`: `repo-intel 1 → 3`, `agnix 1 → 0`.
+- Updated the `/axiom` section in `README.md` to document the new SessionStart hook, `--detect-only` flag, Codex plugin-hooks opt-in (`[features].plugin_hooks = true`), and the OpenCode scaffold limitation pending [sst/opencode#5409](https://github.com/sst/opencode/issues/5409).
+
+### Tests
+- Updated `__tests__/marketplace-standalone-contract.test.js` to pin the `axiom` expected version, ref, and commit at `v0.6.2 / e3f3fab`.
+
 ## [5.13.4] - 2026-05-18
 
 ### Changed

README.md

@@ -19,7 +19,7 @@
 </p>
 
 <p align="center">
-  <b>24 plugins · 49 agents · 45 skills (across all repos) · 30k lines of lib code · 3,513 tests · 5 platforms</b><br>
+  <b>24 plugins · 50 agents · 45 skills (across all repos) · 30k lines of lib code · 3,518 tests · 5 platforms</b><br>
   <em>Plugins distributed as standalone repos under <a href="https://github.com/agent-sh">agent-sh</a> org - agentsys is the marketplace &amp; installer</em>
 </p>
 
@@ -41,11 +41,11 @@
 AI models can write code. That's not the hard part anymore. The hard part is everything around it - task selection, branch management, code review, artifact cleanup, CI, PR comments, deployment. **AgentSys is the runtime that orchestrates agents to handle all of it** - structured pipelines, gated phases, specialized agents, and persistent state that survives session boundaries.
 
 ---
-> Building custom skills, agents, hooks, or MCP tools? [agnix](https://github.com/agent-sh/agnix) is the CLI + LSP linter that catches config errors before they fail silently - real-time IDE validation, auto suggestions, auto-fix, and 399 rules for Claude Code, Codex, OpenCode, Cursor, Kiro, Copilot, Gemini CLI, Cline, Windsurf, Roo Code, Amp, and more.
+> Building custom skills, agents, hooks, or MCP tools? [agnix](https://github.com/agent-sh/agnix) is the CLI + LSP linter that catches config errors before they fail silently - real-time IDE validation, auto suggestions, auto-fix, and 423 rules for Claude Code, Codex, OpenCode, Cursor, Kiro, Copilot, Gemini CLI, Cline, Windsurf, Roo Code, Amp, and more.
 
 ## What This Is
 
-An agent orchestration system - 24 plugins, 49 agents (39 file-based + 10 role-based specialists in audit-project), and 45 skills that compose into structured pipelines for software development. Each plugin lives in its own standalone repo under the [agent-sh](https://github.com/agent-sh) org. agentsys is the marketplace and installer that ties them together.
+An agent orchestration system - 24 plugins, 50 agents (40 file-based + 10 role-based specialists in audit-project), and 45 skills that compose into structured pipelines for software development. Each plugin lives in its own standalone repo under the [agent-sh](https://github.com/agent-sh) org. agentsys is the marketplace and installer that ties them together.
 
 Each agent has a single responsibility, a specific model assignment, and defined inputs/outputs. Pipelines enforce phase gates so agents can't skip steps. State persists across sessions so work survives interruptions.
 
@@ -120,7 +120,7 @@ The investment shifts from model spend to pipeline design. Better prompts, riche
 | [`/gate-and-ship`](#gate-and-ship) | Quality gates then ship (/prepare-delivery + /ship) |
 | [`/axiom`](#axiom) | Durable memory: load, query, list, bootstrap projects, and record approved knowledge |
 | [`/banthis`](#banthis) | Durable negative memory: persist banned agent behaviors |
-| [`/agnix`](#agnix) | Lint agent configurations (399 rules) |
+| [`/agnix`](#agnix) | Lint agent configurations (423 rules) |
 | [`/ship`](#ship) | PR creation, CI monitoring, merge |
 | [`/deslop`](#deslop) | Clean AI slop patterns |
 | [`/perf`](#perf) | Performance investigation with baselines and profiling |
@@ -318,11 +318,14 @@ Each piece runs independently - use `/prepare-delivery` alone to review before d
 
 **[axiom](https://github.com/agent-sh/axiom)** is a standalone plugin and CLI. It creates a private `axiom-based` knowledge repo after explicit approval, keeps only thin context loaded automatically, and stores durable decisions, memories, preferences, and project notes in queryable files.
 
+**Auto-loading at session start (v0.6.2+):** The plugin ships a `SessionStart` hook that runs `axiom before-any --auto-project --detect-only --quiet` automatically in Claude Code and Codex (Codex requires `[features].plugin_hooks = true` in `~/.codex/config.toml`). The hook never mutates state - it emits `## Axiom Setup` when `~/.axiom` is missing or `## Project Detection` when the current git project has no scaffold, then the agent asks the user before running `axiom init` or `axiom project <slug>`. An OpenCode plugin scaffold (`opencode-plugin/axiom.mjs`) hooks `session.created` and pre-warms `~/.axiom/.session-context.md`; full session-start context injection waits on [sst/opencode#5409](https://github.com/sst/opencode/issues/5409). On Cursor / Cline / Aider / Gemini CLI the agent invokes `before-any` via the skill instead.
+
 **What it does:**
 
 | Command | Use |
 |---------|-----|
 | `axiom before-any --quiet` | Load global thin context at the start of meaningful work |
+| `axiom before-any --auto-project --detect-only --quiet` | Read-only auto-load used by the SessionStart hook |
 | `axiom before-any --project <slug>` | Load project context and create missing project scaffolds |
 | `axiom query "<keyword>" --project <slug>` | Retrieve focused, source-backed project knowledge |
 | `axiom list --topics --project <slug>` | Explore what knowledge exists before querying |
@@ -389,7 +392,7 @@ agnix catches these issues before they cause problems.
 | **Best Practices** | Tool restrictions, model selection, trigger phrase quality |
 | **Cross-Platform** | Compatibility across Claude Code, Codex, OpenCode, Cursor, Kiro, Copilot, Gemini CLI, Cline, Windsurf, Roo Code, Amp, and more |
 
-**399 validation rules** (126 auto-fixable) derived from:
+**423 validation rules** (129 auto-fixable) derived from:
 - Official tool specifications (Claude Code, Codex CLI, OpenCode, Cursor, Kiro, GitHub Copilot, Gemini CLI, Cline, Windsurf, Roo Code, Amp, and more)
 - Research papers on agent reliability and prompt injection
 - Real-world testing across 500+ repositories
@@ -1327,7 +1330,7 @@ The system is built on research, not guesswork.
 - Instruction following reliability
 
 **Testing:**
-- 3,513 tests passing
+- 3,518 tests passing
 - Drift-detect validated on 1,000+ repositories
 - E2E workflow testing across all commands
 - Cross-platform validation (Claude Code, OpenCode, Codex CLI, Cursor, Kiro)

__tests__/marketplace-standalone-contract.test.js

@@ -25,9 +25,9 @@ const expectedStandalonePlugins = {
     category: 'development',
   },
   axiom: {
-    version: '0.6.1',
-    ref: 'v0.6.1',
-    commit: 'ba77271a4ff7084d2d567cf3173204390be80f3b',
+    version: '0.6.2',
+    ref: 'v0.6.2',
+    commit: 'e3f3fabfcb19c140a38cbd5c5129ac69ca50b359',
     command: '/axiom',
     category: 'productivity',
   },

docs/reference/AGENTS.md

@@ -3,7 +3,7 @@
 Complete reference for all agents in AgentSys.
 
 <!-- GEN:START:agents-counts -->
-**TL;DR:** 49 agents across 24 plugins (18 have agents). opus for reasoning, sonnet for patterns, haiku for execution. Each agent does one thing well. <!-- AGENT_COUNT_TOTAL: 49 -->
+**TL;DR:** 50 agents across 24 plugins (17 have agents). opus for reasoning, sonnet for patterns, haiku for execution. Each agent does one thing well. <!-- AGENT_COUNT_TOTAL: 50 -->
 <!-- GEN:END:agents-counts -->
 
 ---
@@ -24,7 +24,7 @@ Complete reference for all agents in AgentSys.
 
 ## Overview
 
-AgentSys uses 49 specialized agents across 24 plugins (18 have agents; gate-and-ship is commands-only, axiom, banthis, skill-curator, and system-prompt-curator are skill/command-only, and zig-lsp is a config-only LSP plugin with no commands or agents). Each agent is optimized for a specific task and assigned a model based on complexity:
+AgentSys uses 50 specialized agents across 24 plugins (17 have agents; gate-and-ship is commands-only, axiom, banthis, skill-curator, system-prompt-curator, and agnix are skill/command-only, and zig-lsp is a config-only LSP plugin with no commands or agents). Each agent is optimized for a specific task and assigned a model based on complexity:
 
 | Model | Use Case | Cost |
 |-------|----------|------|
@@ -33,7 +33,7 @@ AgentSys uses 49 specialized agents across 24 plugins (18 have agents; gate-and-
 | haiku | Mechanical execution, no judgment | Low |
 
 **Agent types:**
-- **File-based agents** (39) - Defined in `plugins/*/agents/*.md` with frontmatter <!-- AGENT_COUNT_FILE_BASED: 39 -->
+- **File-based agents** (40) - Defined in `plugins/*/agents/*.md` with frontmatter <!-- AGENT_COUNT_FILE_BASED: 40 -->
 - **Role-based agents** (10) - Defined inline via Task tool with specialized prompts <!-- AGENT_COUNT_ROLE_BASED: 10 -->
 
 ---

package.json

@@ -1,6 +1,6 @@
 {
   "name": "agentsys",
-  "version": "5.13.4",
+  "version": "5.13.5",
   "description": "A modular runtime and orchestration system for AI agents - works with Claude Code, OpenCode, and Codex CLI",
   "main": "lib/platform/detect-platform.js",
   "type": "commonjs",

scripts/generate-docs.js

@@ -445,12 +445,12 @@ const STATIC_PLUGIN_AGENT_COUNTS = {
   'drift-detect': 1,
   'enhance': 8,
   'sync-docs': 1,
-  'repo-intel': 1,
+  'repo-intel': 3,
   'axiom': 0,
   'banthis': 0,
   'perf': 6,
   'learn': 1,
-  'agnix': 1,
+  'agnix': 0,
   'consult': 1,
   'debate': 1,
   'web-ctl': 1,