cc-compact: recursively summarize the /clear -> cc-compact lineage

Follow the chain of ancestor sessions each session was compacted from, up
to --max-depth (default 10), summarizing each older session more tightly
via a per-depth decay factor. Guards against cycles and already-visited
sessions. Bounds the full chain to ~24k tokens worst case.

.dotter/global.toml

@@ -65,6 +65,7 @@ depends = []
 
 [claude.files]
 "claude/CLAUDE.md" = "~/.claude/CLAUDE.md"
+"claude/settings.json" = "~/.claude/settings.json"
 
 [claude.variables]
 

.gitignore

@@ -7,3 +7,10 @@
 # Mudpuppy skills (not tracked)
 claude/skills/mudpuppy-implementation-review/
 claude/skills/mudpuppy-pr-review/
+
+# Python bytecode cache
+__pycache__/
+*.pyc
+
+# htop runtime history
+htop/htop_history

CLAUDE.md

@@ -1,5 +1,9 @@
 # Dotfiles Configuration Notes
 
+## Commit Directly to Main
+
+Commit directly to `main` in this repo. It reflects the current state of my machine, so feature branches don't make sense here. Do not create branches or open PRs for changes.
+
 ## Dotter Template Handling
 
 Dotter automatically treats files with `{{` as templates. For files that contain `{{` in their code (like Lua tables), disable templating by setting `type = "symbolic"` in `.dotter/global.toml`.

claude/CLAUDE.md

@@ -1,5 +1,14 @@
 # Personal Claude Code Instructions
 
+## Do not modify my machine without permission
+
+**Never install software, packages, or tools on my machine without explicit permission** — no `brew install`, `npm install -g`, `pip install`, `apt`, downloading binaries, or anything that changes global system state. This is absolutely disallowed. If a task needs a tool I don't have, stop and ask first, or find a way that doesn't touch my machine (e.g. run it in Docker — if Docker is not available then ask me).
+
+## Agents
+
+If work is parallel or cleanly scoped out, consider handing off to subagents to preserve context.
+Do **not** use fork subagents. Only separate agents with clear scope or a clear goal.
+
 ## Notifying me / sharing images via Discord
 
 The `discord-send` CLI is on my PATH. Use it whenever I ask you to notify me when something is done, or to show me an image or video — send it to my Discord and I'll see it there.
@@ -13,6 +22,14 @@ discord-send --attach ./screenshot.png 'Here is the rendered page'
 
 When asked to respond to reviews in a PR, check the PR description and all comments. Check out the PR branch. Decide if any changes are required or not, make changes if needed, commit and push once done. Always push the PR branch if responding to a PR, do not push to a different branch. 
 
+## Commits
+Commit often. It's better to commit things than to leave changes hanging for a long time. You don't have to commit after every change, but once you're done working just go ahead and commit. 
+
+## Comments in code
+Only comment what is necessary to explain **why** the code is doing something. Do not write comments for **what** the code is doing, that should be obvious from the code itself.
+Do not write comments for anything obvious, anything that an expert can be expected to know, or anything already documented elsewhere.
+Comments are permanent part of the code! Only comment what should be permanently stored along with the code, moment to moment coding decisions, what old versions of the code were before refactors etc. do not belong in comments.
+
 ## Writing Style
 
 Following rules are for your writing. Follow these when responding to the user, when writing documentation and code comments, and when writing commit messages or PR descriptions.
@@ -28,11 +45,13 @@ Following rules are for your writing. Follow these when responding to the user,
 - Do not use em-dashes. No "It's not X, it's Y". Avoid overusing intricate, underscore, highlight, showcase, journey to, step into. Do not use "You're absolutely right!".
 - Avoid sentence fragments for emphasis. Bad: "Not X. Not Y. Just Z."
 - Do use emphasis when necessary: "This is a destructive operation. All data will be lost. Should I proceed?"
+- Avoid flowery language like "no ceremony", "this is more than X", "you're right to poke it", "this is load bearing", "dissolves your original concern"
 
 Examples:
 - Bad: Sure! I'd be happy to help you with that. The issue you're experiencing is likely caused by a problem with the authentication middleware. I will investigate this problem now.
 - Bad: Not a bug. Not a feature. A fundamental design flaw.
 - Bad: It's worth noting that this approach has limitations.
+- Bad: Just implementation, no ceremony.
 - Good: Sure. Bug may be in auth middleware. I'll check.
 - Good: This is a design flaw.
 - Good: This approach has limitations.

claude/settings.json

@@ -0,0 +1,79 @@
+{
+  "attribution": {
+    "sessionUrl": false
+  },
+  "includeCoAuthoredBy": false,
+  "permissions": {
+    "allow": [
+      "mcp__claude_ai_Willet__list_organizations",
+      "mcp__claude_ai_Willet__list_organization_members",
+      "mcp__claude_ai_Willet__list_projects",
+      "mcp__claude_ai_Willet__list_project_members",
+      "mcp__claude_ai_Willet__list_tasks",
+      "mcp__claude_ai_Willet__list_tags",
+      "mcp__claude_ai_Willet__search_tasks",
+      "mcp__claude_ai_Willet__get_task",
+      "mcp__claude_ai_Willet__get_task_graph",
+      "mcp__claude_ai_Willet__get_project_stats",
+      "mcp__claude_ai_Willet__render_dependency_graph",
+      "mcp__claude_ai_Willet__render_task_board",
+      "mcp__claude_ai_Willet__create_task",
+      "mcp__claude_ai_Willet__update_task",
+      "mcp__claude_ai_Willet__complete_task",
+      "mcp__claude_ai_Willet__cancel_task",
+      "mcp__claude_ai_Willet__reopen_task",
+      "mcp__claude_ai_Willet__start_task",
+      "mcp__claude_ai_Willet__link_tasks",
+      "mcp__claude_ai_Willet__unlink_tasks",
+      "mcp__claude_ai_Willet__add_comment"
+    ],
+    "defaultMode": "auto"
+  },
+  "worktree": {
+    "baseRef": "fresh"
+  },
+  "enableWorkflows": false,
+  "workflowKeywordTriggerEnabled": false,
+  "statusLine": {
+    "type": "command",
+    "command": "~/.claude/ccline/ccline",
+    "padding": 0
+  },
+  "enabledPlugins": {
+    "frontend-design@claude-plugins-official": true,
+    "honeycomb@honeycomb-plugins": true,
+    "caveman@caveman": false
+  },
+  "extraKnownMarketplaces": {
+    "mcp-apps": {
+      "source": {
+        "source": "github",
+        "repo": "modelcontextprotocol/ext-apps"
+      }
+    },
+    "honeycomb-plugins": {
+      "source": {
+        "source": "github",
+        "repo": "honeycombio/agent-skill"
+      }
+    },
+    "caveman": {
+      "source": {
+        "source": "github",
+        "repo": "JuliusBrussee/caveman"
+      }
+    }
+  },
+  "alwaysThinkingEnabled": true,
+  "effortLevel": "medium",
+  "promptSuggestionEnabled": false,
+  "awaySummaryEnabled": false,
+  "showClearContextOnPlanAccept": true,
+  "tui": "fullscreen",
+  "autoMemoryEnabled": false,
+  "editorMode": "vim",
+  "fileCheckpointingEnabled": false,
+  "remoteControlAtStartup": true,
+  "agentPushNotifEnabled": true,
+  "skipAutoPermissionPrompt": true
+}

claude/skills/cc-compact/SKILL.md

@@ -16,8 +16,13 @@ The helper script lives at `~/.claude/skills/cc-compact/compact_session.py`.
 
 ## Step 1: Resolve the session
 
-The argument is one of two forms:
+The argument is one of three forms:
 
+- **Nothing** — no id or title given. This is the common case: the user ran
+  `/clear` and then `cc-compact` to reload the session they just cleared. Run
+  the script with no selector; it auto-picks the most recently active session
+  in the current project, **excluding this session** (the fresh one `/clear`
+  created). This emulates Claude Code's built-in `/compact`.
 - **A session id** — when invoked like `/resume claude --resume <id>` or
   `cc-compact <id>`, the user already gave you the UUID. Pass it as `--id`.
 - **A session name / title** — free text. Pass it as `--title`; the script
@@ -27,6 +32,9 @@ The argument is one of two forms:
 Run the script exactly once with the matching form:
 
 ```sh
+# Latest session before the /clear (default, no selector):
+python3 ~/.claude/skills/cc-compact/compact_session.py
+
 # By id:
 python3 ~/.claude/skills/cc-compact/compact_session.py --id <session-uuid>
 
@@ -36,7 +44,8 @@ python3 ~/.claude/skills/cc-compact/compact_session.py --title "the session name
 
 ## Step 2: Read the output
 
-It prints a bounded, XML-tagged report:
+Everything is wrapped in a `<compacted-session-chain>` element containing one or
+more `<session depth="N">` blocks. Each block is a bounded, XML-tagged report:
 
 - header: project, git branch, time span, record/prompt counts
 - the first few exchanges (user prompt + the agent's reply) — the original intent
@@ -46,6 +55,14 @@ It prints a bounded, XML-tagged report:
 - the last several tool calls — what the agent was physically doing last
 - the final assistant message — what it was saying / waiting on last
 
+`depth="0"` (`role="requested"`) is the session you asked for. If that session
+was itself resumed from an earlier one via `/clear` → cc-compact, the script
+follows that lineage backwards and emits each older session at `depth="1"`, `2`,
+… (`role="ancestor"`), summarized more tightly at each step so the total stays
+bounded. It walks up to `--max-depth` ancestors (default 10) and guards against
+cycles. Read the deeper blocks as fading background: the further back, the
+terser. You do not need to do anything to trigger this — it happens on its own.
+
 ## Step 3: Pick up the work
 
 The output is context for *you* — treat it like a resumed session, not something

claude/skills/cc-compact/compact_session.py

@@ -18,6 +18,10 @@ Session resolution (pick one):
   --id UUID          find <UUID>.jsonl under the projects dir
   --title TEXT       find the session whose ai-title contains TEXT
                      (case-insensitive substring; newest match wins)
+  --latest           the most recently active session in the current project,
+                     excluding the caller's own session — this is the default
+                     when no selector is given, so `cc-compact` right after
+                     `/clear` picks up the session you just cleared
 
 Sessions live at: ~/.claude/projects/<encoded-cwd>/<session-id>.jsonl
 """
@@ -27,6 +31,7 @@ import glob
 import json
 import os
 import random
+import re
 import sys
 
 PROJECTS_DIR = os.path.expanduser("~/.claude/projects")
@@ -84,7 +89,87 @@ def is_genuine_prompt(rec):
     return not s.startswith(skip_prefixes)
 
 
+def current_session():
+    """(project_dir, session_id) for the session invoking this script.
+
+    Claude Code exports CLAUDE_CODE_SESSION_ID; its log lives at
+    <project_dir>/<id>.jsonl. Fall back to the encoded cwd when the env var is
+    missing (script run by hand outside a session). Either value may be None."""
+    sid = os.environ.get("CLAUDE_CODE_SESSION_ID")
+    if sid:
+        matches = glob.glob(os.path.join(PROJECTS_DIR, "**", f"{sid}.jsonl"), recursive=True)
+        if matches:
+            return os.path.dirname(matches[0]), sid
+    # Claude Code encodes the cwd into the project dir name by replacing every
+    # non-alphanumeric char with a dash (/Users/kaan/Code/Veery -> -Users-kaan-Code-Veery).
+    encoded = re.sub(r"[^A-Za-z0-9]", "-", os.getcwd())
+    cand = os.path.join(PROJECTS_DIR, encoded)
+    return (cand if os.path.isdir(cand) else None), sid
+
+
+def resolve_latest():
+    """Newest session log in the current project, excluding the caller's own
+    session, so compacting right after `/clear` lands on the just-cleared one."""
+    project_dir, sid = current_session()
+    if project_dir:
+        pool = glob.glob(os.path.join(project_dir, "*.jsonl"))
+    else:
+        pool = glob.glob(os.path.join(PROJECTS_DIR, "**", "*.jsonl"), recursive=True)
+    if sid:
+        pool = [p for p in pool if os.path.basename(p) != f"{sid}.jsonl"]
+    if not pool:
+        sys.exit("No previous session found to compact in this project")
+    latest = max(pool, key=os.path.getmtime)
+    sys.stderr.write(f"Auto-selected latest session: {latest}\n")
+    return latest
+
+
+# A compacted session's log records the cc-compact run it did on its own
+# predecessor: the helper command (carrying --id/--file) and the report it
+# printed (whose <file> tag and "Auto-selected" line hold the resolved absolute
+# path). Following those references walks the /clear -> cc-compact lineage back.
+_ANCESTOR_PATH_RES = [
+    re.compile(r"Auto-selected latest session:\s*(\S+\.jsonl)"),
+    re.compile(r"<file>([^<]+\.jsonl)</file>"),
+    re.compile(r"compact_session\.py[^\n\"]*?--file[= ]+(\S+\.jsonl)"),
+]
+_ANCESTOR_ID_RE = re.compile(r"compact_session\.py[^\n\"]*?--id[= ]+([0-9a-fA-F-]{36})")
+
+
+def find_ancestor(path, visited):
+    """Return the session log `path` was compacted from, or None.
+
+    Scans for references to another session and returns the first that resolves
+    to an existing file not already in `visited` (guards against cycles)."""
+    self_real = os.path.realpath(path)
+    try:
+        fh = open(path, encoding="utf-8", errors="replace")
+    except OSError:
+        return None
+    with fh:
+        for line in fh:
+            if "compact_session.py" not in line and "<file>" not in line and "Auto-selected" not in line:
+                continue
+            candidates = []
+            for pat in _ANCESTOR_PATH_RES:
+                candidates += pat.findall(line)
+            for sid in _ANCESTOR_ID_RE.findall(line):
+                candidates += glob.glob(os.path.join(PROJECTS_DIR, "**", f"{sid}.jsonl"), recursive=True)
+            for cand in candidates:
+                cand = os.path.expanduser(cand)
+                if not os.path.isfile(cand):
+                    continue
+                real = os.path.realpath(cand)
+                if real == self_real or real in visited:
+                    continue
+                return cand
+    return None
+
+
 def resolve_file(args):
+    if args.latest:
+        return resolve_latest()
+
     if args.file:
         return os.path.expanduser(args.file)
 
@@ -131,20 +216,83 @@ def resolve_file(args):
 
 def main():
     ap = argparse.ArgumentParser(description=__doc__, formatter_class=argparse.RawDescriptionHelpFormatter)
-    g = ap.add_mutually_exclusive_group(required=True)
+    g = ap.add_mutually_exclusive_group()
     g.add_argument("--file", help="path to a session .jsonl")
     g.add_argument("--id", help="session UUID")
     g.add_argument("--title", help="substring of the session's ai-title")
+    g.add_argument("--latest", action="store_true",
+                   help="newest session in the current project, excluding the caller's own "
+                        "(the default when no selector is given)")
     ap.add_argument("--first", type=int, default=5, help="how many opening exchanges")
     ap.add_argument("--last", type=int, default=10, help="how many closing exchanges")
     ap.add_argument("--middle", type=int, default=5, help="how many exchanges randomly sampled from the middle")
     ap.add_argument("--top-files", type=int, default=10, help="how many most-edited files")
     ap.add_argument("--maxlen", type=int, default=800, help="max chars per quoted message")
     ap.add_argument("--seed", type=int, default=1, help="seed for the middle random sample")
+    ap.add_argument("--max-depth", type=int, default=10,
+                    help="how many ancestor sessions to follow back through the "
+                         "/clear -> cc-compact chain (0 = just this session)")
+    ap.add_argument("--decay", type=float, default=0.6,
+                    help="per-depth shrink factor for every limit (tighter summaries deeper in the chain)")
     args = ap.parse_args()
+    if not (args.file or args.id or args.title):
+        args.latest = True
 
     path = resolve_file(args)
 
+    out = sys.stdout.write
+
+    # Walk the lineage: the requested session at depth 0, then each older
+    # session it was compacted from, summarized ever more tightly.
+    visited = {os.path.realpath(path)}
+    chain = [path]
+    cur = path
+    for _ in range(max(0, args.max_depth)):
+        anc = find_ancestor(cur, visited)
+        if not anc:
+            break
+        chain.append(anc)
+        visited.add(os.path.realpath(anc))
+        cur = anc
+
+    out(f'<compacted-session-chain sessions="{len(chain)}" '
+        'note="depth 0 is the session you asked for; deeper entries are older '
+        'sessions it was compacted from, summarized more tightly">\n')
+    for depth, p in enumerate(chain):
+        params = scale_params(args, depth)
+        summarize(p, depth, params, out)
+    out("</compacted-session-chain>\n")
+
+
+# Floors keep even the deepest summary useful without letting it grow.
+_FLOORS = {"first": 1, "last": 2, "middle": 0, "top_files": 3, "maxlen": 200, "final": 300, "tools": 3}
+_BASE_FINAL = 2000  # final-agent-message clip at depth 0
+_BASE_TOOLS = 8     # last-tool-calls shown at depth 0
+
+
+def scale_params(args, depth):
+    """Depth-0 uses the CLI limits; each deeper level multiplies every limit by
+    args.decay ** depth, floored so summaries stay non-empty."""
+    f = args.decay ** depth
+
+    def s(base, floor):
+        return max(floor, int(round(base * f)))
+
+    return {
+        "first": s(args.first, _FLOORS["first"]),
+        "last": s(args.last, _FLOORS["last"]),
+        "middle": s(args.middle, _FLOORS["middle"]),
+        "top_files": s(args.top_files, _FLOORS["top_files"]),
+        "maxlen": s(args.maxlen, _FLOORS["maxlen"]),
+        "final": s(_BASE_FINAL, _FLOORS["final"]),
+        "tools": s(_BASE_TOOLS, _FLOORS["tools"]),
+        "seed": args.seed,
+    }
+
+
+def summarize(path, depth, p, out):
+    """Load one session log and emit its bounded, XML-tagged report under a
+    <session depth="N"> element. `p` holds the per-depth limits."""
     turns = []            # conversational turns: {"prompt": str, "reply": [text,...]}
     edit_counts = {}      # file path -> edit count
     last_assistant_text = []  # final text blocks
@@ -203,16 +351,17 @@ def main():
     # sampled only from what's left between them. If the regions collide
     # (short session), the overlap is simply dropped.
     n = len(turns)
-    first_idx = list(range(min(args.first, n)))
-    last_start = max(len(first_idx), n - args.last)
+    first_idx = list(range(min(p["first"], n)))
+    last_start = max(len(first_idx), n - p["last"])
     last_idx = list(range(last_start, n))
     mid_pool = list(range(len(first_idx), last_start))
-    if args.seed is not None:
-        random.seed(args.seed)
-    mid_idx = sorted(random.sample(mid_pool, min(args.middle, len(mid_pool))))
+    if p["seed"] is not None:
+        random.seed(p["seed"])
+    mid_idx = sorted(random.sample(mid_pool, min(p["middle"], len(mid_pool))))
 
     # ---- emit a bounded, XML-tagged report (newlines preserved) ----
-    out = sys.stdout.write
+    role = "requested" if depth == 0 else "ancestor"
+    out(f'\n<session depth="{depth}" role="{role}">\n')
 
     out("<session-summary>\n")
     out(f"  <file>{esc(path)}</file>\n")
@@ -231,9 +380,9 @@ def main():
             t = turns[i]
             reply = "\n".join(t["reply"]).strip()
             out(f'  <exchange n="{i + 1}">\n')
-            out(f"    <user>\n{clip(t['prompt'], args.maxlen)}\n    </user>\n")
+            out(f"    <user>\n{clip(t['prompt'], p['maxlen'])}\n    </user>\n")
             if reply:
-                out(f"    <agent>\n{clip(reply, args.maxlen)}\n    </agent>\n")
+                out(f"    <agent>\n{clip(reply, p['maxlen'])}\n    </agent>\n")
             else:
                 out("    <agent note=\"no text reply — tool calls only\" />\n")
             out("  </exchange>\n")
@@ -243,20 +392,22 @@ def main():
     emit_exchanges("sampled-middle", mid_idx)
     emit_exchanges("last", last_idx)
 
-    out(f'\n<most-edited-files top="{args.top_files}">\n')
-    for fp, c in sorted(edit_counts.items(), key=lambda kv: -kv[1])[: args.top_files]:
+    out(f'\n<most-edited-files top="{p["top_files"]}">\n')
+    for fp, c in sorted(edit_counts.items(), key=lambda kv: -kv[1])[: p["top_files"]]:
         out(f'  <file edits="{c}">{esc(fp)}</file>\n')
     out("</most-edited-files>\n")
 
     out("\n<last-tool-calls>\n")
-    for name, arg in recent_tools[-8:]:
+    for name, arg in recent_tools[-p["tools"]:]:
         out(f'  <call tool="{esc(name)}">{esc(arg)}</call>\n')
     out("</last-tool-calls>\n")
 
     out("\n<final-agent-message>\n")
-    out(clip("\n".join(last_assistant_text), 2000) + "\n")
+    out(clip("\n".join(last_assistant_text), p["final"]) + "\n")
     out("</final-agent-message>\n")
 
+    out("</session>\n")
+
 
 if __name__ == "__main__":
     main()

claude/skills/compact-file/SKILL.md

@@ -0,0 +1,111 @@
+---
+name: compact-file
+description: >
+  Compress natural language memory files (CLAUDE.md, todos, preferences) into shortened format
+  to save input tokens. Preserves all technical substance, code, URLs, and structure.
+  Compressed version overwrites the original file. Human-readable backup saved as FILE.original.md.
+  Trigger: /compact-file FILEPATH or "compact memory file"
+---
+
+# Compact File
+
+## Purpose
+
+Compress natural language files (CLAUDE.md, todos, preferences) into caveman-speak to reduce input tokens. Compressed version overwrites original. Human-readable backup saved as `<filename>.original.md`.
+
+## Trigger
+
+`/compact-file <filepath>` or when user asks to compress a memory file.
+
+## Process
+
+1. The compression scripts live in `scripts/` (adjacent to this SKILL.md). If the path is not immediately available, search for `scripts/__main__.py` next to this SKILL.md.
+
+2. From the directory containing this SKILL.md, run:
+
+python3 -m scripts <absolute_filepath>
+
+3. The CLI will:
+- detect file type (no tokens)
+- call Claude to compress
+- validate output (no tokens)
+- if errors: cherry-pick fix with Claude (targeted fixes only, no recompression)
+- retry up to 2 times
+- if still failing after 2 retries: report error to user, leave original file untouched
+
+4. Return result to user
+
+## Compression Rules
+
+### Remove
+- Articles: a, an, the
+- Filler: just, really, basically, actually, simply, essentially, generally
+- Pleasantries: "sure", "certainly", "of course", "happy to", "I'd recommend"
+- Hedging: "it might be worth", "you could consider", "it would be good to"
+- Redundant phrasing: "in order to" → "to", "make sure to" → "ensure", "the reason is because" → "because"
+- Connective fluff: "however", "furthermore", "additionally", "in addition"
+
+### Preserve EXACTLY (never modify)
+- Code blocks (fenced ``` and indented)
+- Inline code (`backtick content`)
+- URLs and links (full URLs, markdown links)
+- File paths (`/src/components/...`, `./config.yaml`)
+- Commands (`npm install`, `git commit`, `docker build`)
+- Technical terms (library names, API names, protocols, algorithms)
+- Proper nouns (project names, people, companies)
+- Dates, version numbers, numeric values
+- Environment variables (`$HOME`, `NODE_ENV`)
+
+### Preserve Structure
+- All markdown headings (keep exact heading text, compress body below)
+- Bullet point hierarchy (keep nesting level)
+- Numbered lists (keep numbering)
+- Tables (compress cell text, keep structure)
+- Frontmatter/YAML headers in markdown files
+
+### Compress
+- Use short synonyms: "big" not "extensive", "fix" not "implement a solution for", "use" not "utilize"
+- Fragments OK: "Run tests before commit" not "You should always run tests before committing"
+- Drop "you should", "make sure to", "remember to" — just state the action
+- Merge redundant bullets that say the same thing differently
+- Keep one example where multiple examples show the same pattern
+
+CRITICAL RULE:
+Anything inside ``` ... ``` must be copied EXACTLY.
+Do not:
+- remove comments
+- remove spacing
+- reorder lines
+- shorten commands
+- simplify anything
+
+Inline code (`...`) must be preserved EXACTLY.
+Do not modify anything inside backticks.
+
+If file contains code blocks:
+- Treat code blocks as read-only regions
+- Only compress text outside them
+- Do not merge sections around code
+
+## Pattern
+
+Original:
+> You should always make sure to run the test suite before pushing any changes to the main branch. This is important because it helps catch bugs early and prevents broken builds from being deployed to production.
+
+Compressed:
+> Run tests before push to main. Catch bugs early, prevent broken prod deploys.
+
+Original:
+> The application uses a microservices architecture with the following components. The API gateway handles all incoming requests and routes them to the appropriate service. The authentication service is responsible for managing user sessions and JWT tokens.
+
+Compressed:
+> Microservices architecture. API gateway route all requests to services. Auth service manage user sessions + JWT tokens.
+
+## Boundaries
+
+- ONLY compress natural language files (.md, .txt, .typ, .typst, .tex, extensionless)
+- NEVER modify: .py, .js, .ts, .json, .yaml, .yml, .toml, .env, .lock, .css, .html, .xml, .sql, .sh
+- If file has mixed content (prose + code), compress ONLY the prose sections
+- If unsure whether something is code or prose, leave it unchanged
+- Original file is backed up as FILE.original.md before overwriting
+- Never compress FILE.original.md (skip it)

claude/skills/compact-file/scripts/__init__.py

@@ -0,0 +1,9 @@
+"""Caveman compress scripts.
+
+This package provides tools to compress natural language markdown files
+into caveman format to save input tokens.
+"""
+
+__all__ = ["cli", "compress", "detect", "validate"]
+
+__version__ = "1.0.0"

claude/skills/compact-file/scripts/__main__.py

@@ -0,0 +1,3 @@
+from .cli import main
+
+main()

claude/skills/compact-file/scripts/benchmark.py

@@ -0,0 +1,80 @@
+#!/usr/bin/env python3
+from pathlib import Path
+import sys
+
+# Support both direct execution and module import
+try:
+    from .validate import validate
+except ImportError:
+    sys.path.insert(0, str(Path(__file__).parent))
+    from validate import validate
+
+try:
+    import tiktoken
+    _enc = tiktoken.get_encoding("o200k_base")
+except ImportError:
+    _enc = None
+
+
+def count_tokens(text):
+    if _enc is None:
+        return len(text.split())  # fallback: word count
+    return len(_enc.encode(text))
+
+
+def benchmark_pair(orig_path: Path, comp_path: Path):
+    orig_text = orig_path.read_text()
+    comp_text = comp_path.read_text()
+
+    orig_tokens = count_tokens(orig_text)
+    comp_tokens = count_tokens(comp_text)
+    saved = 100 * (orig_tokens - comp_tokens) / orig_tokens if orig_tokens > 0 else 0.0
+    result = validate(orig_path, comp_path)
+
+    return (comp_path.name, orig_tokens, comp_tokens, saved, result.is_valid)
+
+
+def print_table(rows):
+    print("\n| File | Original | Compressed | Saved % | Valid |")
+    print("|------|----------|------------|---------|-------|")
+    for r in rows:
+        print(f"| {r[0]} | {r[1]} | {r[2]} | {r[3]:.1f}% | {'✅' if r[4] else '❌'} |")
+
+
+def main():
+    # Direct file pair: python3 benchmark.py original.md compressed.md
+    if len(sys.argv) == 3:
+        orig = Path(sys.argv[1]).resolve()
+        comp = Path(sys.argv[2]).resolve()
+        if not orig.exists():
+            print(f"❌ Not found: {orig}")
+            sys.exit(1)
+        if not comp.exists():
+            print(f"❌ Not found: {comp}")
+            sys.exit(1)
+        print_table([benchmark_pair(orig, comp)])
+        return
+
+    # Glob mode: repo_root/tests/caveman-compress/
+    # __file__ lives at <repo_root>/skills/caveman-compress/scripts/benchmark.py
+    # Walk up four dirs: scripts → caveman-compress → skills → repo_root.
+    tests_dir = Path(__file__).resolve().parents[3] / "tests" / "caveman-compress"
+    if not tests_dir.exists():
+        print(f"❌ Tests dir not found: {tests_dir}")
+        sys.exit(1)
+
+    rows = []
+    for orig in sorted(tests_dir.glob("*.original.md")):
+        comp = orig.with_name(orig.stem.removesuffix(".original") + ".md")
+        if comp.exists():
+            rows.append(benchmark_pair(orig, comp))
+
+    if not rows:
+        print("No compressed file pairs found.")
+        return
+
+    print_table(rows)
+
+
+if __name__ == "__main__":
+    main()

claude/skills/compact-file/scripts/cli.py

@@ -0,0 +1,85 @@
+#!/usr/bin/env python3
+"""
+Caveman Compress CLI
+
+Usage:
+    caveman <filepath>
+"""
+
+import sys
+
+# Force UTF-8 on stdout/stderr before any code can print. Windows consoles
+# default to cp1252 and crash on the ❌ glyphs in error/validation branches,
+# masking the real error and leaving the user with a half-compressed file.
+for _stream in (sys.stdout, sys.stderr):
+    reconfigure = getattr(_stream, "reconfigure", None)
+    if callable(reconfigure):
+        try:
+            reconfigure(encoding="utf-8", errors="replace")
+        except Exception:
+            pass
+
+from pathlib import Path
+
+from .compress import backup_dir_for, compress_file
+from .detect import detect_file_type, should_compress
+
+
+def print_usage():
+    print("Usage: caveman <filepath>")
+
+
+def main():
+    if len(sys.argv) != 2:
+        print_usage()
+        sys.exit(1)
+
+    filepath = Path(sys.argv[1])
+
+    # Check file exists
+    if not filepath.exists():
+        print(f"❌ File not found: {filepath}")
+        sys.exit(1)
+
+    if not filepath.is_file():
+        print(f"❌ Not a file: {filepath}")
+        sys.exit(1)
+
+    filepath = filepath.resolve()
+
+    # Detect file type
+    file_type = detect_file_type(filepath)
+
+    print(f"Detected: {file_type}")
+
+    # Check if compressible
+    if not should_compress(filepath):
+        print("Skipping: file is not natural language (code/config)")
+        sys.exit(0)
+
+    print("Starting caveman compression...\n")
+
+    try:
+        success = compress_file(filepath)
+
+        if success:
+            print("\nCompression completed successfully")
+            backup_path = backup_dir_for(filepath) / (filepath.stem + ".original.md")
+            print(f"Compressed: {filepath}")
+            print(f"Original:   {backup_path}")
+            sys.exit(0)
+        else:
+            print("\n❌ Compression failed after retries")
+            sys.exit(2)
+
+    except KeyboardInterrupt:
+        print("\nInterrupted by user")
+        sys.exit(130)
+
+    except Exception as e:
+        print(f"\n❌ Error: {e}")
+        sys.exit(1)
+
+
+if __name__ == "__main__":
+    main()

claude/skills/compact-file/scripts/compress.py

@@ -0,0 +1,342 @@
+#!/usr/bin/env python3
+"""
+Caveman Memory Compression Orchestrator
+
+Usage:
+    python scripts/compress.py <filepath>
+"""
+
+import os
+import re
+import shutil
+import subprocess
+import sys
+from pathlib import Path
+from typing import List
+
+OUTER_FENCE_REGEX = re.compile(
+    r"\A\s*(`{3,}|~{3,})[^\n]*\n(.*)\n\1\s*\Z", re.DOTALL
+)
+
+# YAML frontmatter: starts at file start with --- on its own line, ends with --- on its own line.
+# Captures the entire block (including delimiters and trailing newline) and the body after.
+FRONTMATTER_REGEX = re.compile(
+    r"\A(---\r?\n.*?\r?\n---\r?\n)(.*)", re.DOTALL
+)
+
+
+def split_frontmatter(text: str):
+    """Split YAML frontmatter from body. Returns (frontmatter, body).
+
+    Memory files (and many other markdown docs) start with a YAML frontmatter
+    block delimited by `---` lines. The compression LLM has a habit of stripping
+    or rewriting these despite preserve-structure rules in the prompt — so we
+    surgically remove the frontmatter before compression and prepend it back
+    verbatim to the output. Files without frontmatter pass through unchanged.
+    """
+    m = FRONTMATTER_REGEX.match(text)
+    if m:
+        return m.group(1), m.group(2)
+    return "", text
+
+# Filenames and paths that almost certainly hold secrets or PII. Compressing
+# them ships raw bytes to the Anthropic API — a third-party data boundary that
+# developers on sensitive codebases cannot cross. detect.py already skips .env
+# by extension, but credentials.md / secrets.txt / ~/.aws/credentials would
+# slip through the natural-language filter. This is a hard refuse before read.
+SENSITIVE_BASENAME_REGEX = re.compile(
+    r"(?ix)^("
+    r"\.env(\..+)?"
+    r"|\.netrc"
+    r"|credentials(\..+)?"
+    r"|secrets?(\..+)?"
+    r"|passwords?(\..+)?"
+    r"|id_(rsa|dsa|ecdsa|ed25519)(\.pub)?"
+    r"|authorized_keys"
+    r"|known_hosts"
+    r"|.*\.(pem|key|p12|pfx|crt|cer|jks|keystore|asc|gpg)"
+    r")$"
+)
+
+SENSITIVE_PATH_COMPONENTS = frozenset({".ssh", ".aws", ".gnupg", ".kube", ".docker"})
+
+SENSITIVE_NAME_TOKENS = (
+    "secret", "credential", "password", "passwd",
+    "apikey", "accesskey", "token", "privatekey",
+)
+
+
+def backup_dir_for(filepath: Path) -> Path:
+    """Resolve the out-of-tree backup directory for a given source file.
+
+    Backups must live OUTSIDE the source directory so skill auto-loaders
+    (Claude Code rules/, opencode instructions/, etc.) stop re-ingesting the
+    `.original.md` copies as live files. Base dir is platform-aware:
+      - Windows: %LOCALAPPDATA%\\caveman-compress\\backups
+      - else:    $XDG_DATA_HOME/caveman-compress/backups if set,
+                 else ~/.local/share/caveman-compress/backups
+
+    The source file's parent-dir name is mirrored under the base to reduce
+    cross-project collisions (e.g. two `task.md` files in different repos).
+    """
+    if os.name == "nt" or sys.platform == "win32":
+        local_appdata = os.environ.get("LOCALAPPDATA")
+        base = Path(local_appdata) if local_appdata else Path.home() / "AppData" / "Local"
+        base = base / "caveman-compress" / "backups"
+    else:
+        xdg = os.environ.get("XDG_DATA_HOME")
+        base = Path(xdg) if xdg else Path.home() / ".local" / "share"
+        base = base / "caveman-compress" / "backups"
+    return base / filepath.parent.name
+
+
+def is_sensitive_path(filepath: Path) -> bool:
+    """Heuristic denylist for files that must never be shipped to a third-party API."""
+    name = filepath.name
+    if SENSITIVE_BASENAME_REGEX.match(name):
+        return True
+    lowered_parts = {p.lower() for p in filepath.parts}
+    if lowered_parts & SENSITIVE_PATH_COMPONENTS:
+        return True
+    # Normalize separators so "api-key" and "api_key" both match "apikey".
+    lower = re.sub(r"[_\-\s.]", "", name.lower())
+    return any(tok in lower for tok in SENSITIVE_NAME_TOKENS)
+
+
+def strip_llm_wrapper(text: str) -> str:
+    """Strip outer ```markdown ... ``` fence when it wraps the entire output."""
+    m = OUTER_FENCE_REGEX.match(text)
+    if m:
+        return m.group(2)
+    return text
+
+from .detect import should_compress
+from .validate import validate
+
+MAX_RETRIES = 2
+
+
+# ---------- Claude Calls ----------
+
+
+def call_claude(prompt: str) -> str:
+    """Send a prompt to Claude.
+
+    Prefers the Anthropic SDK when ANTHROPIC_API_KEY is set; otherwise falls
+    back to the ``claude --print`` CLI (which handles desktop auth).
+
+    On Windows the CLI subprocess decoding defaults to the system codepage
+    (cp1251 / cp1252) and crashes on UTF-8 output — see issue #152. Pinning
+    ``encoding="utf-8"`` with ``errors="replace"`` matches the CLI's actual
+    native I/O and prevents the UnicodeDecodeError before validation can
+    report. Windows users with non-ASCII content can also set
+    ``ANTHROPIC_API_KEY`` to route through the SDK and skip the subprocess.
+    """
+    api_key = os.environ.get("ANTHROPIC_API_KEY")
+    if api_key:
+        try:
+            import anthropic
+
+            client = anthropic.Anthropic(api_key=api_key)
+            msg = client.messages.create(
+                model=os.environ.get("CAVEMAN_MODEL", "claude-sonnet-4-5"),
+                max_tokens=8192,
+                messages=[{"role": "user", "content": prompt}],
+            )
+            return strip_llm_wrapper(msg.content[0].text.strip())
+        except ImportError:
+            pass  # anthropic not installed, fall back to CLI
+    # Fallback: use claude CLI (handles desktop auth).
+    # Resolve binary via shutil.which so Windows .cmd/.bat shims (e.g.
+    # %APPDATA%\npm\claude.CMD) work without shell=True. On POSIX,
+    # shutil.which returns the same absolute path as the implicit lookup,
+    # so this is a no-op there. Falls back to bare "claude" if not found
+    # on PATH so subprocess raises a clear FileNotFoundError.
+    claude_bin = shutil.which("claude") or "claude"
+    try:
+        result = subprocess.run(
+            [claude_bin, "--print"],
+            input=prompt,
+            text=True,
+            capture_output=True,
+            check=True,
+            encoding="utf-8",
+            errors="replace",
+        )
+        return strip_llm_wrapper(result.stdout.strip())
+    except subprocess.CalledProcessError as e:
+        raise RuntimeError(f"Claude call failed:\n{e.stderr}")
+
+
+def build_compress_prompt(original: str) -> str:
+    return f"""
+Compress this markdown into caveman format.
+
+STRICT RULES:
+- Do NOT modify anything inside ``` code blocks
+- Do NOT modify anything inside inline backticks
+- Preserve ALL URLs exactly
+- Preserve ALL headings exactly
+- Preserve file paths and commands
+- Return ONLY the compressed markdown body — do NOT wrap the entire output in a ```markdown fence or any other fence. Inner code blocks from the original stay as-is; do not add a new outer fence around the whole file.
+
+Only compress natural language.
+
+TEXT:
+{original}
+"""
+
+
+def build_fix_prompt(original: str, compressed: str, errors: List[str]) -> str:
+    errors_str = "\n".join(f"- {e}" for e in errors)
+    return f"""You are fixing a caveman-compressed markdown file. Specific validation errors were found.
+
+CRITICAL RULES:
+- DO NOT recompress or rephrase the file
+- ONLY fix the listed errors — leave everything else exactly as-is
+- The ORIGINAL is provided as reference only (to restore missing content)
+- Preserve caveman style in all untouched sections
+
+ERRORS TO FIX:
+{errors_str}
+
+HOW TO FIX:
+- Missing URL: find it in ORIGINAL, restore it exactly where it belongs in COMPRESSED
+- Code block mismatch: find the exact code block in ORIGINAL, restore it in COMPRESSED
+- Heading mismatch: restore the exact heading text from ORIGINAL into COMPRESSED
+- Do not touch any section not mentioned in the errors
+
+ORIGINAL (reference only):
+{original}
+
+COMPRESSED (fix this):
+{compressed}
+
+Return ONLY the fixed compressed file. No explanation.
+"""
+
+
+# ---------- Core Logic ----------
+
+
+def compress_file(filepath: Path) -> bool:
+    # Resolve and validate path
+    filepath = filepath.resolve()
+    MAX_FILE_SIZE = 500_000  # 500KB
+    if not filepath.exists():
+        raise FileNotFoundError(f"File not found: {filepath}")
+    if filepath.stat().st_size > MAX_FILE_SIZE:
+        raise ValueError(f"File too large to compress safely (max 500KB): {filepath}")
+
+    # Refuse files that look like they contain secrets or PII. Compressing ships
+    # the raw bytes to the Anthropic API — a third-party boundary — so we fail
+    # loudly rather than silently exfiltrate credentials or keys. Override is
+    # intentional: the user must rename the file if the heuristic is wrong.
+    if is_sensitive_path(filepath):
+        raise ValueError(
+            f"Refusing to compress {filepath}: filename looks sensitive "
+            "(credentials, keys, secrets, or known private paths). "
+            "Compression sends file contents to the Anthropic API. "
+            "Rename the file if this is a false positive."
+        )
+
+    print(f"Processing: {filepath}")
+
+    if not should_compress(filepath):
+        print("Skipping (not natural language)")
+        return False
+
+    original_text = filepath.read_text(errors="ignore")
+    # Store backup outside the source directory so skill auto-loaders don't
+    # re-ingest the `.original.md` copy as a live file. Mirror the source's
+    # parent-dir name + stem under a platform-aware base to reduce collisions.
+    backup_dir = backup_dir_for(filepath)
+    backup_dir.mkdir(parents=True, exist_ok=True)
+    backup_path = backup_dir / (filepath.stem + ".original.md")
+
+    if not original_text.strip():
+        print("❌ Refusing to compress: file is empty or whitespace-only.")
+        return False
+
+    # Check if backup already exists to prevent accidental overwriting
+    if backup_path.exists():
+        print(f"⚠️ Backup file already exists: {backup_path}")
+        print("The original backup may contain important content.")
+        print("Aborting to prevent data loss. Please remove or rename the backup file if you want to proceed.")
+        return False
+
+    # Split YAML frontmatter off before compression. Claude tends to strip or
+    # rewrite frontmatter despite preserve-structure rules; we keep it verbatim
+    # by removing it from the input and re-prepending it to the output.
+    frontmatter, body = split_frontmatter(original_text)
+    if frontmatter:
+        print(f"Detected YAML frontmatter ({len(frontmatter)} chars) — preserving verbatim")
+
+    if not body.strip():
+        print("❌ Refusing to compress: body is empty after frontmatter removal.")
+        return False
+
+    # Step 1: Compress (body only, frontmatter excluded)
+    print("Compressing with Claude...")
+    compressed_body = call_claude(build_compress_prompt(body))
+
+    if compressed_body is None or not compressed_body.strip():
+        print("❌ Compression aborted: Claude returned an empty response.")
+        print("   Original file is untouched (no backup created).")
+        return False
+
+    # Compare the BODY (not the whole file) — frontmatter is preserved verbatim
+    # and would never change, so identity must be judged on the compressible part.
+    if compressed_body.strip() == body.strip():
+        print("❌ Compression aborted: output is identical to input.")
+        print("   Likely causes: Claude refused, returned the prompt verbatim, or the file is")
+        print("   already in caveman form. Original file is untouched (no backup created).")
+        return False
+
+    # Reassemble: frontmatter (verbatim) + compressed body
+    compressed = frontmatter + compressed_body
+
+    # Save original as backup, then verify the backup readback before
+    # touching the input file. If the filesystem dropped bytes (encoding,
+    # antivirus, disk full), unlink the bad backup and abort instead of
+    # leaving the user with a corrupt backup + compressed primary.
+    backup_path.write_text(original_text)
+    backup_readback = backup_path.read_text(errors="ignore")
+    if backup_readback != original_text:
+        print(f"❌ Backup write verification failed: {backup_path}")
+        print("   In-memory original differs from on-disk backup. Aborting before touching the input file.")
+        try:
+            backup_path.unlink()
+        except OSError:
+            pass
+        return False
+    filepath.write_text(compressed)
+
+    # Step 2: Validate + Retry
+    for attempt in range(MAX_RETRIES):
+        print(f"\nValidation attempt {attempt + 1}")
+
+        result = validate(backup_path, filepath)
+
+        if result.is_valid:
+            print("Validation passed")
+            break
+
+        print("❌ Validation failed:")
+        for err in result.errors:
+            print(f"   - {err}")
+
+        if attempt == MAX_RETRIES - 1:
+            # Restore original on failure
+            filepath.write_text(original_text)
+            backup_path.unlink(missing_ok=True)
+            print("❌ Failed after retries — original restored")
+            return False
+
+        print("Fixing with Claude...")
+        compressed = call_claude(
+            build_fix_prompt(original_text, compressed, result.errors)
+        )
+        filepath.write_text(compressed)
+
+    return True

claude/skills/compact-file/scripts/detect.py

@@ -0,0 +1,121 @@
+#!/usr/bin/env python3
+"""Detect whether a file is natural language (compressible) or code/config (skip)."""
+
+import json
+import re
+from pathlib import Path
+
+# Extensions that are natural language and compressible
+COMPRESSIBLE_EXTENSIONS = {".md", ".txt", ".markdown", ".rst", ".typ", ".typst", ".tex"}
+
+# Extensions that are code/config and should be skipped
+SKIP_EXTENSIONS = {
+    ".py", ".js", ".ts", ".tsx", ".jsx", ".json", ".yaml", ".yml",
+    ".toml", ".env", ".lock", ".css", ".scss", ".html", ".xml",
+    ".sql", ".sh", ".bash", ".zsh", ".go", ".rs", ".java", ".c",
+    ".cpp", ".h", ".hpp", ".rb", ".php", ".swift", ".kt", ".lua",
+    ".dockerfile", ".makefile", ".csv", ".ini", ".cfg",
+}
+
+# Patterns that indicate a line is code
+CODE_PATTERNS = [
+    re.compile(r"^\s*(import |from .+ import |require\(|const |let |var )"),
+    re.compile(r"^\s*(def |class |function |async function |export )"),
+    re.compile(r"^\s*(if\s*\(|for\s*\(|while\s*\(|switch\s*\(|try\s*\{)"),
+    re.compile(r"^\s*[\}\]\);]+\s*$"),  # closing braces/brackets
+    re.compile(r"^\s*@\w+"),  # decorators/annotations
+    re.compile(r'^\s*"[^"]+"\s*:\s*'),  # JSON-like key-value
+    re.compile(r"^\s*\w+\s*=\s*[{\[\(\"']"),  # assignment with literal
+]
+
+
+def _is_code_line(line: str) -> bool:
+    """Check if a line looks like code."""
+    return any(p.match(line) for p in CODE_PATTERNS)
+
+
+def _is_json_content(text: str) -> bool:
+    """Check if content is valid JSON."""
+    try:
+        json.loads(text)
+        return True
+    except (json.JSONDecodeError, ValueError):
+        return False
+
+
+def _is_yaml_content(lines: list[str]) -> bool:
+    """Heuristic: check if content looks like YAML."""
+    yaml_indicators = 0
+    for line in lines[:30]:
+        stripped = line.strip()
+        if stripped.startswith("---"):
+            yaml_indicators += 1
+        elif re.match(r"^\w[\w\s]*:\s", stripped):
+            yaml_indicators += 1
+        elif stripped.startswith("- ") and ":" in stripped:
+            yaml_indicators += 1
+    # If most non-empty lines look like YAML
+    non_empty = sum(1 for l in lines[:30] if l.strip())
+    return non_empty > 0 and yaml_indicators / non_empty > 0.6
+
+
+def detect_file_type(filepath: Path) -> str:
+    """Classify a file as 'natural_language', 'code', 'config', or 'unknown'.
+
+    Returns:
+        One of: 'natural_language', 'code', 'config', 'unknown'
+    """
+    ext = filepath.suffix.lower()
+
+    # Extension-based classification
+    if ext in COMPRESSIBLE_EXTENSIONS:
+        return "natural_language"
+    if ext in SKIP_EXTENSIONS:
+        return "code" if ext not in {".json", ".yaml", ".yml", ".toml", ".ini", ".cfg", ".env"} else "config"
+
+    # Extensionless files (like CLAUDE.md, TODO) — check content
+    if not ext:
+        try:
+            text = filepath.read_text(errors="ignore")
+        except (OSError, PermissionError):
+            return "unknown"
+
+        lines = text.splitlines()[:50]
+
+        if _is_json_content(text[:10000]):
+            return "config"
+        if _is_yaml_content(lines):
+            return "config"
+
+        code_lines = sum(1 for l in lines if l.strip() and _is_code_line(l))
+        non_empty = sum(1 for l in lines if l.strip())
+        if non_empty > 0 and code_lines / non_empty > 0.4:
+            return "code"
+
+        return "natural_language"
+
+    return "unknown"
+
+
+def should_compress(filepath: Path) -> bool:
+    """Return True if the file is natural language and should be compressed."""
+    if not filepath.is_file():
+        return False
+    # Skip backup files
+    if filepath.name.endswith(".original.md"):
+        return False
+    return detect_file_type(filepath) == "natural_language"
+
+
+if __name__ == "__main__":
+    import sys
+
+    if len(sys.argv) < 2:
+        print("Usage: python detect.py <file1> [file2] ...")
+        sys.exit(1)
+
+    for path_str in sys.argv[1:]:
+        p = Path(path_str).resolve()
+        file_type = detect_file_type(p)
+        compress = should_compress(p)
+        print(f"  {p.name:30s} type={file_type:20s} compress={compress}")

claude/skills/compact-file/scripts/validate.py

@@ -0,0 +1,213 @@
+#!/usr/bin/env python3
+import re
+from collections import Counter
+from pathlib import Path
+
+URL_REGEX = re.compile(r"https?://[^\s)]+")
+FENCE_OPEN_REGEX = re.compile(r"^(\s{0,3})(`{3,}|~{3,})(.*)$")
+HEADING_REGEX = re.compile(r"^(#{1,6})\s+(.*)", re.MULTILINE)
+BULLET_REGEX = re.compile(r"^\s*[-*+]\s+", re.MULTILINE)
+
+# crude but effective path detection
+# Requires either a path prefix (./ ../ / or drive letter) or a slash/backslash within the match
+PATH_REGEX = re.compile(r"(?:\./|\.\./|/|[A-Za-z]:\\)[\w\-/\\\.]+|[\w\-\.]+[/\\][\w\-/\\\.]+")
+
+
+class ValidationResult:
+    def __init__(self):
+        self.is_valid = True
+        self.errors = []
+        self.warnings = []
+
+    def add_error(self, msg):
+        self.is_valid = False
+        self.errors.append(msg)
+
+    def add_warning(self, msg):
+        self.warnings.append(msg)
+
+
+def read_file(path: Path) -> str:
+    return path.read_text(errors="ignore")
+
+
+# ---------- Extractors ----------
+
+
+def extract_headings(text):
+    return [(level, title.strip()) for level, title in HEADING_REGEX.findall(text)]
+
+
+def extract_code_blocks(text):
+    """Line-based fenced code block extractor.
+
+    Handles ``` and ~~~ fences with variable length (CommonMark: closing
+    fence must use same char and be at least as long as opening). Supports
+    nested fences (e.g. an outer 4-backtick block wrapping inner 3-backtick
+    content).
+    """
+    blocks = []
+    lines = text.split("\n")
+    i = 0
+    n = len(lines)
+    while i < n:
+        m = FENCE_OPEN_REGEX.match(lines[i])
+        if not m:
+            i += 1
+            continue
+        fence_char = m.group(2)[0]
+        fence_len = len(m.group(2))
+        open_line = lines[i]
+        block_lines = [open_line]
+        i += 1
+        closed = False
+        while i < n:
+            close_m = FENCE_OPEN_REGEX.match(lines[i])
+            if (
+                close_m
+                and close_m.group(2)[0] == fence_char
+                and len(close_m.group(2)) >= fence_len
+                and close_m.group(3).strip() == ""
+            ):
+                block_lines.append(lines[i])
+                closed = True
+                i += 1
+                break
+            block_lines.append(lines[i])
+            i += 1
+        if closed:
+            blocks.append("\n".join(block_lines))
+        # Unclosed fences are silently skipped — they indicate malformed markdown
+        # and including them would cause false-positive validation failures.
+    return blocks
+
+
+def extract_urls(text):
+    return set(URL_REGEX.findall(text))
+
+
+def extract_paths(text):
+    return set(PATH_REGEX.findall(text))
+
+
+def count_bullets(text):
+    return len(BULLET_REGEX.findall(text))
+
+
+def extract_inline_codes(text):
+    text_without_fences = re.sub(r"^```[\s\S]*?^```", "", text, flags=re.MULTILINE)
+    text_without_fences = re.sub(r"^~~~[\s\S]*?^~~~", "", text_without_fences, flags=re.MULTILINE)
+    return re.findall(r"`([^`]+)`", text_without_fences)
+
+
+# ---------- Validators ----------
+
+
+def validate_headings(orig, comp, result):
+    h1 = extract_headings(orig)
+    h2 = extract_headings(comp)
+
+    if len(h1) != len(h2):
+        result.add_error(f"Heading count mismatch: {len(h1)} vs {len(h2)}")
+
+    if h1 != h2:
+        result.add_warning("Heading text/order changed")
+
+
+def validate_code_blocks(orig, comp, result):
+    c1 = extract_code_blocks(orig)
+    c2 = extract_code_blocks(comp)
+
+    if c1 != c2:
+        result.add_error("Code blocks not preserved exactly")
+
+
+def validate_urls(orig, comp, result):
+    u1 = extract_urls(orig)
+    u2 = extract_urls(comp)
+
+    if u1 != u2:
+        result.add_error(f"URL mismatch: lost={u1 - u2}, added={u2 - u1}")
+
+
+def validate_paths(orig, comp, result):
+    p1 = extract_paths(orig)
+    p2 = extract_paths(comp)
+
+    if p1 != p2:
+        result.add_warning(f"Path mismatch: lost={p1 - p2}, added={p2 - p1}")
+
+
+def validate_bullets(orig, comp, result):
+    b1 = count_bullets(orig)
+    b2 = count_bullets(comp)
+
+    if b1 == 0:
+        return
+
+    diff = abs(b1 - b2) / b1
+
+    if diff > 0.15:
+        result.add_warning(f"Bullet count changed too much: {b1} -> {b2}")
+
+
+def validate_inline_codes(orig, comp, result):
+    c1 = Counter(extract_inline_codes(orig))
+    c2 = Counter(extract_inline_codes(comp))
+
+    if c1 != c2:
+        lost = set(c1.keys()) - set(c2.keys())
+        added = set(c2.keys()) - set(c1.keys())
+        for code, count in c1.items():
+            if code in c2 and c2[code] < count:
+                lost.add(f"{code} (lost {count - c2[code]} of {count} occurrences)")
+        if lost:
+            result.add_error(f"Inline code lost: {lost}")
+        if added:
+            result.add_warning(f"Inline code added: {added}")
+
+
+# ---------- Main ----------
+
+
+def validate(original_path: Path, compressed_path: Path) -> ValidationResult:
+    result = ValidationResult()
+
+    orig = read_file(original_path)
+    comp = read_file(compressed_path)
+
+    validate_headings(orig, comp, result)
+    validate_code_blocks(orig, comp, result)
+    validate_urls(orig, comp, result)
+    validate_paths(orig, comp, result)
+    validate_bullets(orig, comp, result)
+    validate_inline_codes(orig, comp, result)
+
+    return result
+
+
+# ---------- CLI ----------
+
+if __name__ == "__main__":
+    import sys
+
+    if len(sys.argv) != 3:
+        print("Usage: python validate.py <original> <compressed>")
+        sys.exit(1)
+
+    orig = Path(sys.argv[1]).resolve()
+    comp = Path(sys.argv[2]).resolve()
+
+    res = validate(orig, comp)
+
+    print(f"\nValid: {res.is_valid}")
+
+    if res.errors:
+        print("\nErrors:")
+        for e in res.errors:
+            print(f"  - {e}")
+
+    if res.warnings:
+        print("\nWarnings:")
+        for w in res.warnings:
+            print(f"  - {w}")

claude/skills/new-work/SKILL.md

@@ -16,9 +16,10 @@ If there are uncommitted changes:
 
 ## Step 2: Switch to main and pull latest
 
-- Attempt to check out `main`.
-  - If that fails because the branch does not exist, check out `master` instead.
-- Then pull the latest changes:
+- Attempt to check out `main`, or failing that `master`.
+- Then pull the latest changes.
+
+You **must** do this. Unless instructed to continue on the same branch, make sure you start from latest main.
 
 ## Step 3: Create a new branch
 

claude/skills/sqlite/SKILL.md

@@ -0,0 +1,32 @@
+---
+name: sqlite
+description: Recommended SQLite defaults (PRAGMAs and STRICT tables). Use when setting up a new project that uses SQLite, or when adding new SQLite tables to an existing project.
+---
+
+Apply these when starting a new SQLite project or adding tables. They are the sane defaults SQLite ships without.
+
+## Connection PRAGMAs
+
+Run these on **every new connection** (except `journal_mode`, see below). They are per-connection state, not stored in the DB file. If your library has a "connection init" hook, put them there.
+
+```sql
+PRAGMA journal_mode = WAL;      -- set once, persists in the DB file
+PRAGMA busy_timeout = 5000;     -- wait 5s on a locked DB instead of erroring
+PRAGMA synchronous = NORMAL;    -- safe to lower from FULL only under WAL
+PRAGMA foreign_keys = ON;       -- SQLite ignores FK constraints unless you enable this
+```
+
+## Tables
+
+Create tables as `STRICT` so column types are enforced:
+
+```sql
+CREATE TABLE user (id INTEGER PRIMARY KEY, name TEXT NOT NULL) STRICT;
+```
+
+## Gotchas
+
+- **`foreign_keys`, `busy_timeout`, `synchronous` reset per connection.** Missing them on any connection silently drops the behavior (e.g. FK constraints stop being enforced). Set them on connection open.
+- **`journal_mode = WAL` is different**: it persists in the DB file, so set it once. WAL needs the DB on **local disk** (breaks on network filesystems) and adds `-wal` and `-shm` sidecar files next to the DB.
+- **`synchronous = NORMAL` durability**: an OS crash or power loss can lose the last few committed transactions. The DB stays consistent (no corruption) — only the tail is lost. `NORMAL` is almost always the right choice; `FULL` is rarely needed. If you genuinely cannot afford to lose any committed write, you shouldn't be on SQLite at all — use a DB replicated across regions.
+- **STRICT allows only `INT`, `INTEGER`, `REAL`, `TEXT`, `BLOB`, `ANY`.** No `VARCHAR`, `DATETIME`, `BOOLEAN`, etc. Use `ANY` for a column that needs to hold mixed types.

git/gitignore

@@ -1,2 +1,6 @@
 .history/
 
+
+## Panda
+styled-system
+styled-system-studio

htop/htop_history

@@ -1 +0,0 @@
-python