@softspark/ai-toolkit 1.5.0 → 1.6.0

package/CHANGELOG.md

@@ -7,6 +7,31 @@ Versioning follows [Semantic Versioning](https://semver.org/).
 
 ---
 
+## v1.6.0 — IDE Language Rules Propagation, Planning Docs, Cloud Security Pack (2026-04-10)
+
+### Added
+- **Language rules propagation to all IDE editors** — shared `dir_rules_shared.py` module now injects language-specific and registered rules into Cursor, Windsurf, Cline, Roo Code, Augment, Antigravity, and Copilot generators. All platforms receive identical rule content from a single source of truth.
+- **Local Dashboard Plan** — planning doc (`kb/planning/local-dashboard-plan.md`) for ai-toolkit UI with visual skill/agent management features
+- **Enterprise Config Inheritance Plan** — planning doc (`kb/planning/enterprise-config-inheritance-plan.md`) for hierarchical config system
+- **Offline SLM Profile Plan** — planning doc (`kb/planning/offline-slm-profile-plan.md`) for offline small language model profiles
+- **Cloud Security Pack Plan** — planning doc (`kb/planning/cloud-security-pack-plan.md`) for multi-cloud audit (GCP/AWS/Azure)
+- **193 new generator tests** — language rules propagation, content verification, cross-platform parity
+
+### Changed
+- **Documentation standards** — added `planning` as a valid KB category in `validate.py`, `documenter` agent, `/docs` and `/documentation-standards` skills
+- **Maintenance SOP** — updated to reflect language rules propagation workflow
+
+---
+
+## v1.5.1 — Security Hardening: Script Injection, XSS, Private Data Leak (2026-04-10)
+
+### Fixed
+- **`action.yml` script injection** — replaced `${{ inputs.command }}` direct interpolation with `env:` variable to prevent GitHub Actions script injection (OWASP A03)
+- **`visual-server.cjs` stored XSS** — extracted inline script to `poll.js`, added `Content-Security-Policy: script-src 'self'` header to block injected scripts in PRD visual preview
+- **`strip_private.py` regex** — changed `[^<]*` to `.*?` with `re.DOTALL` flag to correctly handle multi-line `<private>` blocks and inner angle brackets (was leaking private data in edge cases)
+
+---
+
 ## v1.5.0 — HIPAA Scanner: Deterministic Script, CI Integration, Self-Exclusion (2026-04-10)
 
 ### Added

package/README.md

@@ -6,7 +6,7 @@
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](LICENSE)
 [![Skills](https://img.shields.io/badge/skills-92-brightgreen)](app/skills/)
 [![Agents](https://img.shields.io/badge/agents-44-blue)](app/agents/)
-[![Tests](https://img.shields.io/badge/tests-439%20passing-success)](tests/)
+[![Tests](https://img.shields.io/badge/tests-451%20passing-success)](tests/)
 
 ---
 
@@ -128,7 +128,7 @@ Replaces all symlinks with real files, inlines rules into CLAUDE.md, copies cons
 | Google Antigravity | `.agent/rules/*.md` + `.agent/workflows/*.md` | `ai-toolkit install --local` | project |
 | Codex / OpenCode | `AGENTS.md` | `ai-toolkit agents-md` | project |
 
-> **Note:** Claude Code is always installed (primary platform with full feature support). Other editors are installed on demand with `--editors <list>` or auto-detected from existing project files. All platforms receive the same agent/skill catalog, guidelines, and rules. For editors lacking native bash lifecycle hooks, `--local` installs a Git hooks fallback (`.git/hooks/pre-commit`) to enforce quality gates pre-commit.
+> **Note:** Claude Code is always installed (primary platform with full feature support). Other editors are installed on demand with `--editors <list>` or auto-detected from existing project files. All platforms receive the same agent/skill catalog, guidelines, rules, language-specific rules, and registered custom rules. For editors lacking native bash lifecycle hooks, `--local` installs a Git hooks fallback (`.git/hooks/pre-commit`) to enforce quality gates pre-commit.
 
 ---
 
@@ -191,7 +191,7 @@ ai-toolkit/
 │   ├── output-styles/   # System prompt output style overrides
 │   ├── constitution.md  # 5 immutable safety articles
 │   └── ARCHITECTURE.md  # Full system design
-├── kb/                  # Reference docs, architecture notes, procedures
+├── kb/                  # Reference docs, architecture notes, procedures, plans
 ├── scripts/             # Validation, install, evaluation scripts
 ├── tests/               # Bats test suite
 └── CHANGELOG.md
@@ -529,6 +529,19 @@ ai-toolkit install --local --lang go,python   # multiple languages
 
 `--local` automatically detects languages using two-phase detection: config markers (package.json, go.mod, Cargo.toml, etc.) plus source file extension scanning (.py, .ts, .go, etc.). `--lang` accepts aliases (`go`, `c++`, `cs`). Rules are injected into `CLAUDE.md` and auto-updated on `ai-toolkit update --local`.
 
+When `--editors` is used alongside detected or explicit languages, language rules are propagated to all configured editors — not just Claude. Each editor receives the full rule content in its native format:
+
+| Editor | Language rule file | Activation |
+|--------|-------------------|------------|
+| Cursor | `.cursor/rules/ai-toolkit-lang-<lang>.mdc` | `globs` per file type (e.g. `**/*.py`) |
+| Windsurf | `.windsurf/rules/ai-toolkit-lang-<lang>.md` | always loaded |
+| Cline | `.cline/rules/ai-toolkit-lang-<lang>.md` | always loaded |
+| Roo Code | `.roo/rules/ai-toolkit-lang-<lang>.md` | always loaded |
+| Augment | `.augment/rules/ai-toolkit-lang-<lang>.md` | `auto_attached` with globs |
+| Antigravity | `.agent/rules/ai-toolkit-lang-<lang>.md` | always loaded |
+
+Registered rules (`ai-toolkit add-rule`) are also propagated to directory-based editor configs as `ai-toolkit-custom-<name>` files.
+
 ---
 
 ## Extension API

package/action.yml

@@ -39,8 +39,9 @@ runs:
     - name: Run ai-toolkit
       id: run-toolkit
       shell: bash
+      env:
+        COMMAND: ${{ inputs.command }}
       run: |
-        COMMAND="${{ inputs.command }}"
         case "$COMMAND" in
           validate|doctor) ;;
           *) echo "Error: invalid command '$COMMAND' (allowed: validate, doctor)"; exit 1 ;;

package/app/.claude-plugin/plugin.json

@@ -1,7 +1,7 @@
 {
   "name": "ai-toolkit",
   "description": "Professional-grade Claude Code toolkit with persona presets, skill security auditor, expanded lifecycle hooks, experimental opt-in plugin packs, benchmark harvesting, and multi-tool support.",
-  "version": "1.5.0",
+  "version": "1.6.0",
   "author": {
     "name": "SoftSpark",
     "url": "https://github.com/softspark"

package/app/ARCHITECTURE.md

@@ -305,7 +305,9 @@ Lead Session (You)
 `app/plugins/mcp-templates/` ships 25 ready-to-use MCP server config templates (filesystems, databases, GitHub, Slack, etc.). Opt-in via `ai-toolkit install --modules mcp-templates` or activated automatically with `--profile strict|full`.
 
 ### Language Rules (70 files, 13 languages)
-`app/rules/` contains per-language coding rules injected into `CLAUDE.md`. Supported languages: TypeScript, Python, Go, Rust, Java, Kotlin, Swift, Dart, C#, PHP, C++, Ruby, and common (shared). Auto-detected from project files via `--auto-detect` or selected with `--modules rules-<lang>`.
+`app/rules/` contains per-language coding rules. Supported languages: TypeScript, Python, Go, Rust, Java, Kotlin, Swift, Dart, C#, PHP, C++, Ruby, and common (shared). Auto-detected from project files via `--auto-detect` or selected with `--modules rules-<lang>`.
+
+Language rules are propagated to **all configured editors** — not just Claude. `dir_rules_shared.build_language_rules()` reads `app/rules/<lang>/*.md`, strips frontmatter, and returns combined content per language. Each directory-based generator (Cursor, Windsurf, Cline, Roo, Augment, Antigravity) emits `ai-toolkit-lang-<lang>` files in its native format. Registered custom rules (`~/.ai-toolkit/rules/`) are similarly propagated as `ai-toolkit-custom-<name>` files via `build_registered_rules()`.
 
 ### Extension API (`inject-hook`)
 The `inject_section_cli.py` script provides a stable marker-based injection API. Any tool can add sections to `CLAUDE.md`, `constitution.md`, or `ARCHITECTURE.md` without overwriting user content, using `<!-- TOOLKIT:START:<id> -->` / `<!-- TOOLKIT:END:<id> -->` markers.

package/app/agents/documenter.md

@@ -202,12 +202,12 @@ command
 
 ## Frontmatter Standards (MANDATORY)
 
-Follow the `documentation-standards` knowledge skill for full spec. **7 required fields:** title, category, service, tags, created, last_updated, description. **5 valid categories:** reference, howto, procedures, troubleshooting, best-practices. `validate.sh` enforces compliance — **docs without valid frontmatter block CI.**
+Follow the `documentation-standards` knowledge skill for full spec. **7 required fields:** title, category, service, tags, created, last_updated, description. **6 valid categories:** reference, howto, procedures, troubleshooting, best-practices, planning. `validate.sh` enforces compliance — **docs without valid frontmatter block CI.**
 
 ## Hard Rules (ENFORCED — NO EXCEPTIONS)
 
 1. **REFUSE** to create any file in `kb/` without valid YAML frontmatter containing ALL 7 required fields (title, category, service, tags, created, last_updated, description).
-2. **REFUSE** to use any category other than: `reference`, `howto`, `procedures`, `troubleshooting`, `best-practices`.
+2. **REFUSE** to use any category other than: `reference`, `howto`, `procedures`, `troubleshooting`, `best-practices`, `planning`.
 3. **REFUSE** to place a document in a directory that doesn't match its `category:` field (e.g., a `howto` doc MUST go in `kb/howto/`).
 4. **REFUSE** to write KB content in any language other than English.
 5. **ALWAYS** run `ai-toolkit validate` or `scripts/validate.py` after creating/modifying KB documents to verify compliance.

package/app/plugins/memory-pack/scripts/strip_private.py

@@ -14,7 +14,7 @@ import sys
 
 def main() -> None:
     content = sys.stdin.read()
-    cleaned = re.sub(r"<private>[^<]*</private>", "", content)
+    cleaned = re.sub(r"<private>.*?</private>", "", content, flags=re.DOTALL)
     sys.stdout.write(cleaned)
 
 

package/app/skills/docs/SKILL.md

@@ -44,7 +44,7 @@ Generate or update documentation.
 When creating or updating documents in the `kb/` directory, follow the `documentation-standards` knowledge skill (auto-loaded). Key rules:
 
 - **7 required frontmatter fields:** title, category, service, tags, created, last_updated, description
-- **5 valid categories:** reference, howto, procedures, troubleshooting, best-practices
+- **6 valid categories:** reference, howto, procedures, troubleshooting, best-practices, planning
 - **Category must match directory** (e.g., `category: howto` → `kb/howto/`)
 - **English only.** No exceptions.
 - **validate.sh rejects docs without valid frontmatter.**

package/app/skills/documentation-standards/SKILL.md

@@ -178,7 +178,7 @@ scripts/validate.py
 # Checks: required fields present, category is valid, tags non-empty
 ```
 
-Valid categories: `reference`, `howto`, `procedures`, `troubleshooting`, `best-practices`.
+Valid categories: `reference`, `howto`, `procedures`, `troubleshooting`, `best-practices`, `planning`.
 
 ## Anti-Patterns
 

package/app/skills/write-a-prd/scripts/frame-template.html

@@ -82,30 +82,6 @@
     </div>
   </main>
   <div class="status"><span class="dot"></span>Connected</div>
-  <script>
-    (function () {
-      var contentEl = document.getElementById("content");
-      var lastHtml = "";
-
-      function poll() {
-        fetch("/content")
-          .then(function (res) { return res.json(); })
-          .then(function (data) {
-            if (data.html && data.html !== lastHtml) {
-              lastHtml = data.html;
-              contentEl.style.opacity = "0";
-              setTimeout(function () {
-                contentEl.innerHTML = data.html;
-                contentEl.style.opacity = "1";
-              }, 80);
-            }
-          })
-          .catch(function () {});
-      }
-
-      setInterval(poll, 2000);
-      poll();
-    })();
-  </script>
+  <script src="/poll.js"></script>
 </body>
 </html>

package/app/skills/write-a-prd/scripts/poll.js

@@ -0,0 +1,23 @@
+(function () {
+  var contentEl = document.getElementById("content");
+  var lastHtml = "";
+
+  function poll() {
+    fetch("/content")
+      .then(function (res) { return res.json(); })
+      .then(function (data) {
+        if (data.html && data.html !== lastHtml) {
+          lastHtml = data.html;
+          contentEl.style.opacity = "0";
+          setTimeout(function () {
+            contentEl.innerHTML = data.html;
+            contentEl.style.opacity = "1";
+          }, 80);
+        }
+      })
+      .catch(function () {});
+  }
+
+  setInterval(poll, 2000);
+  poll();
+})();

package/app/skills/write-a-prd/scripts/visual-server.cjs

@@ -20,16 +20,26 @@ function resetIdleTimer() {
 
 const templatePath = path.join(__dirname, "frame-template.html");
 const templateHtml = fs.readFileSync(templatePath, "utf-8");
+const pollJs = fs.readFileSync(path.join(__dirname, "poll.js"), "utf-8");
 
 const server = http.createServer((req, res) => {
   resetIdleTimer();
 
   if (req.method === "GET" && (req.url === "/" || req.url === "/index.html")) {
-    res.writeHead(200, { "Content-Type": "text/html; charset=utf-8" });
+    res.writeHead(200, {
+      "Content-Type": "text/html; charset=utf-8",
+      "Content-Security-Policy": "script-src 'self'",
+    });
     res.end(templateHtml);
     return;
   }
 
+  if (req.method === "GET" && req.url === "/poll.js") {
+    res.writeHead(200, { "Content-Type": "application/javascript; charset=utf-8" });
+    res.end(pollJs);
+    return;
+  }
+
   if (req.method === "GET" && req.url === "/content") {
     res.writeHead(200, { "Content-Type": "application/json" });
     res.end(JSON.stringify({ html: currentContent }));