context-mode 1.0.53 → 1.0.54

package/configs/codex/AGENTS.md

@@ -56,3 +56,22 @@ Search results can flood context. Use `ctx_execute(language: "shell", code: "gre
 | `ctx stats` | Call the `stats` MCP tool and display the full output verbatim |
 | `ctx doctor` | Call the `doctor` MCP tool, run the returned shell command, display as checklist |
 | `ctx upgrade` | Call the `upgrade` MCP tool, run the returned shell command, display as checklist |
+
+## Windows notes
+
+**PowerShell cmdlets in shell scripts** — The sandbox executes scripts via bash. PowerShell
+cmdlets (`Format-List`, `Format-Table`, `Get-Culture`, etc.) do not exist in bash and will fail
+with `command not found`. Wrap them with `pwsh -NoProfile -Command "..."` instead.
+
+**Relative paths** — The sandbox CWD is a temp directory, not your project root. Always convert
+any user-supplied path to an absolute path before passing it to `rg`, `grep`, or `find`.
+Ask the user to confirm the absolute path if it is not already known.
+
+**Windows drive letter paths** — The sandbox runs Git Bash / MSYS2, not WSL. Drive letters must
+use the MSYS2 convention, NOT the WSL convention:
+`X:\path` → `/x/path` (lowercase letter, no `/mnt/` prefix).
+Never emit `/mnt/<letter>/` prefixes regardless of which drive the project is on.
+
+**Always quote paths** — If a path contains spaces, bash splits it into separate arguments.
+Always wrap every path in double quotes: `rg "symbol" "$REPO_ROOT/some dir/Source"`.
+This applies to all tools: `rg`, `grep`, `find`, `ls`, etc.

package/configs/kilo/AGENTS.md

@@ -0,0 +1,58 @@
+# context-mode — MANDATORY routing rules
+
+You have context-mode MCP tools available. These rules are NOT optional — they protect your context window from flooding. A single unrouted command can dump 56 KB into context and waste the entire session.
+
+## BLOCKED commands — do NOT attempt these
+
+### curl / wget — BLOCKED
+Any shell command containing `curl` or `wget` will be intercepted and blocked by the context-mode plugin. Do NOT retry.
+Instead use:
+- `context-mode_ctx_fetch_and_index(url, source)` to fetch and index web pages
+- `context-mode_ctx_execute(language: "javascript", code: "const r = await fetch(...)")` to run HTTP calls in sandbox
+
+### Inline HTTP — BLOCKED
+Any shell command containing `fetch('http`, `requests.get(`, `requests.post(`, `http.get(`, or `http.request(` will be intercepted and blocked. Do NOT retry with shell.
+Instead use:
+- `context-mode_ctx_execute(language, code)` to run HTTP calls in sandbox — only stdout enters context
+
+### Direct web fetching — BLOCKED
+Do NOT use any direct URL fetching tool. Use the sandbox equivalent.
+Instead use:
+- `context-mode_ctx_fetch_and_index(url, source)` then `context-mode_ctx_search(queries)` to query the indexed content
+
+## REDIRECTED tools — use sandbox equivalents
+
+### Shell (>20 lines output)
+Shell is ONLY for: `git`, `mkdir`, `rm`, `mv`, `cd`, `ls`, `npm install`, `pip install`, and other short-output commands.
+For everything else, use:
+- `context-mode_ctx_batch_execute(commands, queries)` — run multiple commands + search in ONE call
+- `context-mode_ctx_execute(language: "shell", code: "...")` — run in sandbox, only stdout enters context
+
+### File reading (for analysis)
+If you are reading a file to **edit** it → reading is correct (edit needs content in context).
+If you are reading to **analyze, explore, or summarize** → use `context-mode_ctx_execute_file(path, language, code)` instead. Only your printed summary enters context.
+
+### grep / search (large results)
+Search results can flood context. Use `context-mode_ctx_execute(language: "shell", code: "grep ...")` to run searches in sandbox. Only your printed summary enters context.
+
+## Tool selection hierarchy
+
+1. **GATHER**: `context-mode_ctx_batch_execute(commands, queries)` — Primary tool. Runs all commands, auto-indexes output, returns search results. ONE call replaces 30+ individual calls.
+2. **FOLLOW-UP**: `context-mode_ctx_search(queries: ["q1", "q2", ...])` — Query indexed content. Pass ALL questions as array in ONE call.
+3. **PROCESSING**: `context-mode_ctx_execute(language, code)` | `context-mode_ctx_execute_file(path, language, code)` — Sandbox execution. Only stdout enters context.
+4. **WEB**: `context-mode_ctx_fetch_and_index(url, source)` then `context-mode_ctx_search(queries)` — Fetch, chunk, index, query. Raw HTML never enters context.
+5. **INDEX**: `context-mode_ctx_index(content, source)` — Store content in FTS5 knowledge base for later search.
+
+## Output constraints
+
+- Keep responses under 500 words.
+- Write artifacts (code, configs, PRDs) to FILES — never return them as inline text. Return only: file path + 1-line description.
+- When indexing content, use descriptive source labels so others can `search(source: "label")` later.
+
+## ctx commands
+
+| Command | Action |
+|---------|--------|
+| `ctx stats` | Call the `stats` MCP tool and display the full output verbatim |
+| `ctx doctor` | Call the `doctor` MCP tool, run the returned shell command, display as checklist |
+| `ctx upgrade` | Call the `upgrade` MCP tool, run the returned shell command, display as checklist |

package/configs/kilo/kilo.json

@@ -0,0 +1,10 @@
+{
+  "$schema": "https://app.kilo.ai/config.json",
+  "mcp": {
+    "context-mode": {
+      "type": "local",
+      "command": ["context-mode"]
+    }
+  },
+  "plugin": ["context-mode"]
+}

package/hooks/core/tool-naming.mjs

@@ -20,6 +20,7 @@ const TOOL_PREFIXES = {
   "gemini-cli":     (tool) => `mcp__context-mode__${tool}`,
   "antigravity":    (tool) => `mcp__context-mode__${tool}`,
   "opencode":       (tool) => `context-mode_${tool}`,
+  "kilo":           (tool) => `context-mode_${tool}`,
   "vscode-copilot": (tool) => `context-mode_${tool}`,
   "kiro":           (tool) => `@context-mode/${tool}`,
   "zed":            (tool) => `mcp:context-mode:${tool}`,

package/hooks/pretooluse.mjs

@@ -87,39 +87,44 @@ try {
       writeFileSync(ipPath, JSON.stringify(ip, null, 2) + "\n", "utf-8");
     }
 
-    // 3. Update hook path + matcher in settings.json
+    // 3. Update hook paths + matcher in settings.json for ALL hook types (#187)
+    //    Previously only fixed PreToolUse — SessionStart, PostToolUse, PreCompact,
+    //    UserPromptSubmit paths remained stale after marketplace auto-update.
     const settingsPath = resolve(homedir(), ".claude", "settings.json");
     try {
       const settings = JSON.parse(readFileSync(settingsPath, "utf-8"));
-      const hooks = settings.hooks?.PreToolUse;
-      if (Array.isArray(hooks)) {
-        let changed = false;
-        for (const entry of hooks) {
-          // Fix deprecated Task-only matcher → Agent|Task
-          if (entry.matcher && entry.matcher.includes("Task") && !entry.matcher.includes("Agent")) {
+      const allHooks = settings.hooks || {};
+      let changed = false;
+
+      for (const hookType of Object.keys(allHooks)) {
+        const entries = allHooks[hookType];
+        if (!Array.isArray(entries)) continue;
+
+        for (const entry of entries) {
+          // Fix deprecated Task-only matcher (PreToolUse only)
+          if (hookType === "PreToolUse" && entry.matcher?.includes("Task") && !entry.matcher.includes("Agent")) {
             entry.matcher = entry.matcher.replace("Task", "Agent|Task");
             changed = true;
           }
+          // Rewrite stale context-mode hook paths to point to current version
           for (const h of (entry.hooks || [])) {
-            if (h.command?.includes("pretooluse.mjs") && !h.command.includes(targetDir)) {
-              h.command = "node " + resolve(targetDir, "hooks", "pretooluse.mjs");
-              changed = true;
+            if (h.command && h.command.includes(".mjs") && h.command.includes("context-mode") && !h.command.includes(targetDir)) {
+              // Extract the script filename (e.g., sessionstart.mjs, pretooluse.mjs)
+              const scriptMatch = h.command.match(/([a-z]+\.mjs)\s*"?\s*$/);
+              if (scriptMatch) {
+                h.command = "node " + resolve(targetDir, "hooks", scriptMatch[1]);
+                changed = true;
+              }
             }
           }
         }
-        if (changed) writeFileSync(settingsPath, JSON.stringify(settings, null, 2) + "\n", "utf-8");
       }
+
+      if (changed) writeFileSync(settingsPath, JSON.stringify(settings, null, 2) + "\n", "utf-8");
     } catch { /* skip settings update */ }
 
-    // 4. Nuke stale version dirs (keep only targetDir and current running dir)
-    try {
-      const keepDirs = new Set([basename(targetDir), myDirName]);
-      for (const d of readdirSync(cacheParent)) {
-        if (!keepDirs.has(d)) {
-          try { rmSync(resolve(cacheParent, d), { recursive: true, force: true }); } catch { /* skip */ }
-        }
-      }
-    } catch { /* skip */ }
+    // Old version dirs are cleaned lazily by sessionstart.mjs (age-gated >1h)
+    // to avoid breaking active sessions that still reference them (#181).
 
     writeFileSync(marker, Date.now().toString(), "utf-8");
   }

package/hooks/sessionstart.mjs

@@ -24,7 +24,7 @@ import { writeSessionEventsFile, buildSessionDirective, getSessionEvents, getLat
 import { createSessionLoaders } from "./session-loaders.mjs";
 import { join, dirname } from "node:path";
 import { fileURLToPath } from "node:url";
-import { readFileSync, writeFileSync, unlinkSync } from "node:fs";
+import { readFileSync, writeFileSync, unlinkSync, readdirSync, rmSync, statSync } from "node:fs";
 import { homedir } from "node:os";
 
 // Resolve absolute path for imports (fileURLToPath for Windows compat)
@@ -120,6 +120,30 @@ try {
     }
 
     db.close();
+
+    // Age-gated lazy cleanup of old plugin cache version dirs (#181).
+    // Only delete dirs older than 1 hour to avoid breaking active sessions.
+    try {
+      const pluginRoot = process.env.CLAUDE_PLUGIN_ROOT;
+      if (pluginRoot) {
+        const cacheParentMatch = pluginRoot.match(/^(.*[\\/]plugins[\\/]cache[\\/][^\\/]+[\\/][^\\/]+[\\/])/);
+        if (cacheParentMatch) {
+          const cacheParent = cacheParentMatch[1];
+          const myDir = pluginRoot.replace(cacheParent, "").replace(/[\\/]/g, "");
+          const ONE_HOUR = 3600000;
+          const now = Date.now();
+          for (const d of readdirSync(cacheParent)) {
+            if (d === myDir) continue;
+            try {
+              const st = statSync(join(cacheParent, d));
+              if (now - st.mtimeMs > ONE_HOUR) {
+                rmSync(join(cacheParent, d), { recursive: true, force: true });
+              }
+            } catch { /* skip */ }
+          }
+        }
+      }
+    } catch { /* best effort — never block session start */ }
   }
   // "clear" — no action needed
 } catch (err) {

package/openclaw.plugin.json

@@ -3,7 +3,7 @@
   "name": "Context Mode",
   "kind": "tool",
   "description": "OpenClaw plugin that saves 98% of your context window. Sandboxed code execution in 11 languages, FTS5 knowledge base with BM25 ranking, and intent-driven search.",
-  "version": "1.0.53",
+  "version": "1.0.54",
   "sandbox": {
     "mode": "permissive",
     "filesystem_access": "full",

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "context-mode",
-  "version": "1.0.53",
+  "version": "1.0.54",
   "type": "module",
   "description": "MCP plugin that saves 98% of your context window. Works with Claude Code, Gemini CLI, VS Code Copilot, OpenCode, and Codex CLI. Sandboxed code execution, FTS5 knowledge base, and intent-driven search.",
   "author": "Mert Koseoğlu",