diff --git a/9p b/9p
index f5c0b29..acaaef9 160000
--- a/9p
+++ b/9p
@@ -1 +1 @@
-Subproject commit f5c0b29522e1cc9e5eafac82fc0022b649bd3685
+Subproject commit acaaef997179b6287b14a10d9df401b08a3be11f
diff --git a/core b/core
index 9ba68ff..abe8ab7 160000
--- a/core
+++ b/core
@@ -1 +1 @@
-Subproject commit 9ba68ff35e2ef099ab2ec9a65cd45ff1b3d2ebd6
+Subproject commit abe8ab76ccc4c66a58752fb43141b0b86bf1cdd6
diff --git a/skills b/skills
index defc568..afb0e64 160000
--- a/skills
+++ b/skills
@@ -1 +1 @@
-Subproject commit defc568841d72f316c640624bd8596d9027400df
+Subproject commit afb0e64262300caeb4cf37b00fa1390a637cedc5
diff --git a/tools/file_edit b/tools/file_edit
new file mode 100755
index 0000000..a78330e
--- /dev/null
+++ b/tools/file_edit
@@ -0,0 +1,130 @@
+#!/usr/bin/env python3
+# description: Replace text in a file; must read with file_read first
+# Args: <file_path> <old_string> <new_string> [replace_all]
+
+import sys
+import os
+import hashlib
+import difflib
+import tempfile
+
+def marker_path(file_path):
+    h = hashlib.sha256(file_path.encode()).hexdigest()[:64]
+    sid = os.environ.get('OLLIE_SESSION_ID', '_')
+    return f"/tmp/ollie/{sid}/{h}"
+
+def unified_diff(path, old, new):
+    old_lines = old.splitlines(keepends=True)
+    new_lines = new.splitlines(keepends=True)
+    return ''.join(difflib.unified_diff(old_lines, new_lines, fromfile=path, tofile=path))
+
+def find_exact(content, find):
+    return find if find in content else None
+
+def find_ws_normalized(content, find):
+    nf = ' '.join(find.split())
+    find_lines = find.split('\n')
+    n = len(find_lines)
+    lines = content.split('\n')
+    for i in range(len(lines) - n + 1):
+        block = '\n'.join(lines[i:i+n])
+        if ' '.join(block.split()) == nf:
+            return block
+    return None
+
+def find_indent_flexible(content, find):
+    def deindent(text):
+        ls = text.split('\n')
+        min_indent = float('inf')
+        for l in ls:
+            stripped = l.lstrip(' \t')
+            if stripped:
+                min_indent = min(min_indent, len(l) - len(stripped))
+        if min_indent == float('inf'):
+            return text
+        return '\n'.join(l[min_indent:] if len(l) > min_indent else l.lstrip(' \t') for l in ls)
+    nf = deindent(find)
+    find_lines = find.split('\n')
+    n = len(find_lines)
+    lines = content.split('\n')
+    for i in range(len(lines) - n + 1):
+        block = '\n'.join(lines[i:i+n])
+        if deindent(block) == nf:
+            return block
+    return None
+
+def find_trimmed_boundary(content, find):
+    trimmed = find.strip()
+    if trimmed == find:
+        return None
+    if trimmed in content:
+        return trimmed
+    find_lines = find.split('\n')
+    n = len(find_lines)
+    lines = content.split('\n')
+    for i in range(len(lines) - n + 1):
+        block = '\n'.join(lines[i:i+n])
+        if block.strip() == trimmed:
+            return block
+    return None
+
+def robust_replace(content, old_string, new_string, replace_all):
+    match = (find_exact(content, old_string) or
+             find_ws_normalized(content, old_string) or
+             find_indent_flexible(content, old_string) or
+             find_trimmed_boundary(content, old_string))
+    if match is None:
+        return None, "old_string not found in file"
+    if replace_all:
+        return content.replace(match, new_string), None
+    n = content.count(match)
+    if n > 1:
+        return None, f"old_string is not unique ({n} occurrences); use replace_all=true or provide more context"
+    return content.replace(match, new_string, 1), None
+
+if len(sys.argv) < 4:
+    print("error: file_path, old_string, and new_string are required")
+    sys.exit(1)
+
+file_path   = sys.argv[1]
+old_string  = sys.argv[2]
+new_string  = sys.argv[3]
+replace_all = len(sys.argv) > 4 and sys.argv[4].lower() in ('true', '1', 'yes')
+
+if not os.path.isabs(file_path):
+    print(f"error: path must be absolute, got: {file_path}")
+    sys.exit(1)
+if old_string == new_string:
+    print("error: old_string and new_string must be different")
+    sys.exit(1)
+if not os.path.exists(file_path):
+    print(f"error: file not found: {file_path}")
+    sys.exit(1)
+
+marker = marker_path(file_path)
+if not os.path.exists(marker):
+    print(f"error: {file_path} must be read first with file_read")
+    sys.exit(1)
+
+with open(file_path, 'r', errors='replace') as f:
+    original = f.read()
+
+new_content, err = robust_replace(original, old_string, new_string, replace_all)
+if err:
+    print(f"error: {err}")
+    sys.exit(1)
+
+tmp = tempfile.NamedTemporaryFile(mode='w', dir=os.path.dirname(file_path) or '.', delete=False)
+try:
+    tmp.write(new_content)
+    tmp.close()
+    os.replace(tmp.name, file_path)
+except Exception as e:
+    os.unlink(tmp.name)
+    print(f"error: {e}")
+    sys.exit(1)
+
+os.unlink(marker)
+
+diff = unified_diff(file_path, original, new_content)
+print(diff if diff else "(no changes)", end='' if diff else '\n')
diff --git a/tools/file_glob b/tools/file_glob
new file mode 100755
index 0000000..9dfb69c
--- /dev/null
+++ b/tools/file_glob
@@ -0,0 +1,14 @@
+#!/usr/bin/env bash
+# description: Find files by name/path pattern, sorted by modification time (newest first)
+# Args: <pattern> [path]
+
+pattern="$1"
+search_path="${2:-$PWD}"
+
+[ -z "$pattern" ] && echo "error: pattern is required" && exit 1
+
+# collect matches and sort by mtime descending
+rg --no-config --files -g "$pattern" "$search_path" 2>/dev/null \
+    | xargs -d '\n' stat --printf '%Y\t%n\n' 2>/dev/null \
+    | sort -rn \
+    | cut -f2-
diff --git a/tools/file_grep b/tools/file_grep
new file mode 100755
index 0000000..4627781
--- /dev/null
+++ b/tools/file_grep
@@ -0,0 +1,88 @@
+#!/usr/bin/env python3
+# description: Search file contents using ripgrep
+# Args: <pattern> [--path=<dir>] [--mode=files_with_matches|content|count] [--glob=<pattern>] [--type=<type>] [-i] [-n] [-A <n>] [-B <n>] [-C <n>] [--multiline] [--head=<n>] [--offset=<n>]
+
+import sys
+import os
+import subprocess
+import argparse
+
+parser = argparse.ArgumentParser(prog='file_grep', add_help=False)
+parser.add_argument('pattern')
+parser.add_argument('--path', default='')
+parser.add_argument('--mode', default='files_with_matches',
+                    choices=['files_with_matches', 'content', 'count'])
+parser.add_argument('--glob', default='')
+parser.add_argument('--type', default='')
+parser.add_argument('-i', action='store_true')
+parser.add_argument('-n', action='store_true', default=True)
+parser.add_argument('-A', type=int, default=None)
+parser.add_argument('-B', type=int, default=None)
+parser.add_argument('-C', type=int, default=None)
+parser.add_argument('--multiline', action='store_true')
+parser.add_argument('--head', type=int, default=0)
+parser.add_argument('--offset', type=int, default=0)
+
+args = parser.parse_args()
+
+rg = 'rg'
+search_path = args.path or os.getcwd()
+
+cmd = ['rg', '--no-config', '--color=never']
+
+if args.mode == 'files_with_matches':
+    cmd.append('-l')
+elif args.mode == 'count':
+    cmd.append('--count')
+elif args.mode == 'content':
+    if args.n:
+        cmd.append('-n')
+
+if args.i:
+    cmd.append('-i')
+if args.multiline:
+    cmd += ['-U', '--multiline-dotall']
+if args.glob:
+    cmd += ['-g', args.glob]
+if args.type:
+    cmd += ['-t', args.type]
+if args.C is not None:
+    cmd += ['-C', str(args.C)]
+if args.A is not None:
+    cmd += ['-A', str(args.A)]
+if args.B is not None:
+    cmd += ['-B', str(args.B)]
+
+cmd += ['-e', args.pattern, search_path]
+
+result = subprocess.run(cmd, capture_output=True, text=True)
+if result.returncode == 1:
+    sys.exit(0)  # no matches
+if result.returncode > 1:
+    print(result.stderr.strip() or f"rg exited with {result.returncode}", file=sys.stderr)
+    sys.exit(1)
+
+lines = result.stdout.splitlines()
+
+if args.offset:
+    lines = lines[args.offset:]
+if args.head:
+    lines = lines[:args.head]
+
+if args.mode == 'content':
+    formatted = []
+    for line in lines:
+        if line == '--':
+            formatted.append(line)
+            continue
+        # file:lineno:content or file-lineno-content (context lines)
+        for sep in (':', '-'):
+            parts = line.split(sep, 2)
+            if len(parts) == 3 and parts[1].isdigit():
+                formatted.append(f"{parts[0]}:{parts[1]:>5}| {parts[2]}")
+                break
+        else:
+            formatted.append(line)
+    lines = formatted
+
+print('\n'.join(lines))
diff --git a/tools/file_read b/tools/file_read
new file mode 100755
index 0000000..5c17b1a
--- /dev/null
+++ b/tools/file_read
@@ -0,0 +1,36 @@
+#!/usr/bin/env bash
+# description: Read a file from the local filesystem with optional offset and line limit
+# Args: <file_path> [offset] [limit]
+
+file_path="$1"
+offset="${2:-1}"
+limit="${3:-2000}"
+max_line_len=2000
+max_size=$((8 * 1024 * 1024))
+
+[ -z "$file_path" ] && echo "error: file_path is required" && exit 1
+[[ "$file_path" != /* ]] && echo "error: path must be absolute, got: $file_path" && exit 1
+[ ! -e "$file_path" ] && echo "error: file not found: $file_path" && exit 1
+[ ! -f "$file_path" ] && echo "error: not a regular file: $file_path" && exit 1
+
+size=$(wc -c < "$file_path")
+[ "$size" -gt "$max_size" ] && echo "error: file too large ($size bytes, max $max_size)" && exit 1
+
+content=$(awk -v off="$offset" -v lim="$limit" -v maxlen="$max_line_len" '
+    NR >= off && NR < off+lim {
+        if (length($0) > maxlen) print substr($0, 1, maxlen) "..."
+        else print
+    }
+    NR >= off+lim { exit }
+' "$file_path")
+
+if [ -z "$content" ]; then
+    echo "(empty file)"
+else
+    printf '%s\n' "$content"
+fi
+
+# create isread marker
+mkdir -p "/tmp/ollie/$OLLIE_SESSION_ID"
+marker="/tmp/ollie/$OLLIE_SESSION_ID/$(printf '%s' "$file_path" | sha256sum | cut -c1-64)"
+touch "$marker"
diff --git a/tools/file_write b/tools/file_write
new file mode 100755
index 0000000..678b322
--- /dev/null
+++ b/tools/file_write
@@ -0,0 +1,61 @@
+#!/usr/bin/env python3
+# description: Write or overwrite a file; must read with file_read first
+# Args: <file_path> <content>
+
+import sys
+import os
+import hashlib
+import difflib
+import tempfile
+
+def marker_path(file_path):
+    h = hashlib.sha256(file_path.encode()).hexdigest()[:64]
+    sid = os.environ.get('OLLIE_SESSION_ID', '_')
+    return f"/tmp/ollie/{sid}/{h}"
+
+def unified_diff(path, old, new):
+    old_lines = old.splitlines(keepends=True)
+    new_lines = new.splitlines(keepends=True)
+    return ''.join(difflib.unified_diff(old_lines, new_lines, fromfile=path, tofile=path))
+
+if len(sys.argv) < 3:
+    print("error: file_path and content are required")
+    sys.exit(1)
+
+file_path = sys.argv[1]
+content   = sys.argv[2]
+
+if not os.path.isabs(file_path):
+    print(f"error: path must be absolute, got: {file_path}")
+    sys.exit(1)
+
+marker = marker_path(file_path)
+if not os.path.exists(marker):
+    print(f"error: {file_path} must be read first with file_read")
+    sys.exit(1)
+
+exists = os.path.exists(file_path)
+old_content = ""
+if exists:
+    with open(file_path, 'r', errors='replace') as f:
+        old_content = f.read()
+
+os.makedirs(os.path.dirname(file_path) or '.', exist_ok=True)
+tmp = tempfile.NamedTemporaryFile(mode='w', dir=os.path.dirname(file_path) or '.', delete=False)
+try:
+    tmp.write(content)
+    tmp.close()
+    os.replace(tmp.name, file_path)
+except Exception as e:
+    os.unlink(tmp.name)
+    print(f"error: {e}")
+    sys.exit(1)
+
+os.unlink(marker)
+
+if not exists:
+    for line in content.splitlines():
+        print(f"+{line}")
+else:
+    diff = unified_diff(file_path, old_content, content)
+    print(diff if diff else "(no changes)", end='' if diff else '\n')
diff --git a/tools/reasoning_think b/tools/reasoning_think
new file mode 100755
index 0000000..d251141
--- /dev/null
+++ b/tools/reasoning_think
@@ -0,0 +1,6 @@
+#!/usr/bin/env bash
+# description: Internal reasoning scratchpad; think through complex problems before acting
+# Args: <thought>
+
+[ -z "$*" ] && echo "error: thought is required" && exit 1
+exit 0
diff --git a/tools/task_complete.py b/tools/task_complete
similarity index 100%
rename from tools/task_complete.py
rename to tools/task_complete
diff --git a/tools/task_plan.sh b/tools/task_plan
similarity index 100%
rename from tools/task_plan.sh
rename to tools/task_plan