CLAUDE.md

1# CLAUDE.md — caveman
2 
3## README is a product artifact
4 
5README = product front door. Non-technical people read it to decide if caveman worth install. Treat like UI copy.
6 
7**Rules for any README change:**
8 
9- Readable by non-AI-agent users. If you write "SessionStart hook injects system context," invisible to most — translate it.
10- Keep Before/After examples first. That the pitch.
11- Install table always complete + accurate. One broken install command costs real user.
12- What You Get table must sync with actual code. Feature ships or removed → update table.
13- Preserve voice. Caveman speak in README on purpose. "Brain still big." "Cost go down forever." "One rock. That it." — intentional brand. Don't normalize.
14- Benchmark numbers from real runs in `benchmarks/` and `evals/`. Never invent or round. Re-run if doubt.
15- Adding new agent to install table → add detail block in `<details>` section below.
16- Readability check before any README commit: would non-programmer understand + install within 60 seconds?
17 
18---
19 
20## Project overview
21 
22Caveman makes AI coding agents respond in compressed caveman-style prose — cuts 65% output tokens (measured), full technical accuracy. Ships as Claude Code plugin, Codex plugin, Gemini CLI extension, agent rule files for Cursor, Windsurf, Cline, Copilot, 40+ others via `npx skills`.
23 
24---
25 
26## What lives where
27 
28Post-cleanup layout. Sources of truth at the top, distribution mirrors below, build outputs in `dist/`, human docs alongside each skill.
29 
30```
31caveman/
32├── README.md                    # Front door (product pitch)
33├── INSTALL.md                   # Per-agent install commands
34├── CONTRIBUTING.md              # Dev guide
35├── CLAUDE.md                    # This file (maintainer instructions)
36├── AGENTS.md / GEMINI.md        # Autodiscovery files (must stay at root)
37│
38├── install.sh / install.ps1     # 30-line shims → cli/install.js
39│
40├── cli/                         # Unified installer
41│   ├── install.js               # Single source for all 30+ agents (PROVIDERS array)
42│   └── lib/settings.js          # JSONC-tolerant settings.json reader/writer
43│
44├── skills/                      # ALL skills, single source of truth
45│   ├── caveman/{SKILL.md, README.md}
46│   ├── caveman-commit/{SKILL.md, README.md}
47│   ├── caveman-review/{SKILL.md, README.md}
48│   ├── caveman-help/{SKILL.md, README.md}
49│   ├── caveman-stats/{SKILL.md, README.md}
50│   ├── caveman-compress/{SKILL.md, README.md, scripts/}
51│   └── cavecrew/{SKILL.md, README.md}
52│
53├── agents/                      # cavecrew subagents (single source — kept at root for plugin auto-discovery)
54├── commands/                    # Codex/Gemini TOML command stubs (root for plugin auto-discovery)
55│
56├── src/                         # Internal source — not auto-discovered by plugin
57│   ├── hooks/                   # Claude Code hooks (installer reads here)
58│   ├── rules/                   # Auto-activation rule body (single source)
59│   ├── tools/                   # caveman-init.js (per-repo rule writer)
60│   └── mcp-servers/             # caveman-shrink npm-published MCP middleware
61│
62├── .claude-plugin/              # Claude Code plugin manifest (REQUIRED at root)
63├── plugins/caveman/             # Claude Code plugin distribution (CI-mirrored)
64│   ├── skills/                  # ← from skills/
65│   └── agents/                  # ← from agents/
66│
67├── dist/                        # Build artifacts (gitignored)
68│   └── caveman.skill            # ZIP of skills/caveman/, rebuilt by CI
69│
70├── tests/                       # All tests (Node + Python)
71├── benchmarks/                  # Real token measurements through Claude API
72├── evals/                       # Three-arm eval harness
73├── docs/                        # User-facing docs site
74└── .github/workflows/           # CI sync
75```
76 
77---
78 
79## File structure and what owns what
80 
81### Single source of truth files — edit only these
82 
83| File | What it controls |
84|------|-----------------|
85| `skills/caveman/SKILL.md` | Caveman behavior: intensity levels, rules, wenyan mode, auto-clarity, persistence. Only file to edit for behavior changes. |
86| `src/rules/caveman-activate.md` | Always-on auto-activation rule body. Consumed by `src/tools/caveman-init.js` when a user runs `npx caveman --with-init` (per-repo IDE rule files). Edit here, not in any per-agent rule copy. |
87| `src/rules/caveman-openclaw-bootstrap.md` | Marker-fenced bootstrap snippet appended to `~/.openclaw/workspace/SOUL.md` by `cli/lib/openclaw.js`. Drives always-on caveman through the OpenClaw gateway. Must include the SENTINEL `Respond terse like smart caveman` and stay well under OpenClaw's 12K-per-bootstrap-file cap. |
88| `cli/lib/openclaw.js` | OpenClaw install/uninstall helper. Frontmatter merge (`version`, `always: true`), SOUL.md marker append/strip, idempotent. Shared by `cli/install.js` and `src/tools/caveman-init.js`. |
89| `skills/caveman-commit/SKILL.md` | Caveman commit message behavior. Fully independent skill. |
90| `skills/caveman-review/SKILL.md` | Caveman code review behavior. Fully independent skill. |
91| `skills/caveman-help/SKILL.md` | Quick-reference card. One-shot display, not a persistent mode. |
92| `skills/caveman-compress/SKILL.md` | Compress sub-skill behavior. |
93| `skills/cavecrew/SKILL.md` | Cavecrew decision guide — when to delegate to caveman subagents vs vanilla. Edit only here. |
94| `agents/cavecrew-investigator.md` | Read-only locator subagent (haiku). Output contract: `path:line — symbol — note`. |
95| `agents/cavecrew-builder.md` | Surgical 1-2 file editor subagent. Refuses 3+ file scope. |
96| `agents/cavecrew-reviewer.md` | Diff/file reviewer subagent (haiku). One-line findings with severity emoji. |
97| `src/plugins/opencode/plugin.js` | opencode native plugin. ESM Bun module — `session.created` writes flag, `tui.prompt.append` parses slash/natural-language activation and appends per-prompt reinforcement. Reuses `caveman-config.js` via `createRequire`. |
98| `src/plugins/opencode/commands/*.md` | Six opencode slash-command prompt templates (`/caveman`, `/caveman-{commit,review,compress,stats,help}`). |
99 
100### Auto-generated / auto-synced — do not edit directly
101 
102We removed the agent-specific dotdir mirrors at the repo root (`.cursor/`, `.windsurf/`, `.clinerules/`, `.github/copilot-instructions.md`, root `caveman/SKILL.md`). They were never read by the installer — only used to self-apply caveman to this repo when a maintainer opened it in Cursor/Windsurf/Cline. Devs who want caveman in their editor while editing this repo should run `npx caveman --with-init` once (writes per-repo rule files from `src/rules/caveman-activate.md` via `src/tools/caveman-init.js`). For per-user installs through the upstream skills CLI, `npx caveman --only <agent>` runs `npx skills add ... -a <profile>`.
103 
104A handful of dotdir leftovers (`.junie/`, `.kiro/`, `.roo/`, `.agents/`) still hold a stale `cavecrew/SKILL.md` mirror from before the cleanup. They aren't read by anything in the current install path; remove on sight, no migration needed.
105 
106What's left is the Claude Code plugin distribution (required by the plugin loader) and the release ZIP.
107 
108| File | Synced from |
109|------|-------------|
110| `plugins/caveman/skills/caveman/SKILL.md` | `skills/caveman/SKILL.md` |
111| `plugins/caveman/skills/caveman-compress/SKILL.md` (+ `scripts/`) | `skills/caveman-compress/SKILL.md` (+ `scripts/`) |
112| `plugins/caveman/skills/cavecrew/SKILL.md` | `skills/cavecrew/SKILL.md` |
113| `plugins/caveman/agents/cavecrew-*.md` | `agents/cavecrew-*.md` |
114| `dist/caveman.skill` | ZIP of `skills/caveman/` directory (gitignored; rebuilt by CI on release) |
115 
116Skills not in this table (`caveman-commit`, `caveman-review`, `caveman-help`, `caveman-stats`) are not mirrored into the Claude Code plugin distribution by CI. They reach Claude Code through the standalone hook + skill install path, and reach other agents via `npx skills add`. A `plugins/caveman/skills/caveman-stats/` directory is currently checked in as a hand-committed copy; the sync workflow does not touch it, so don't rely on edits there to propagate.
117 
118---
119 
120## CI sync workflow
121 
122`.github/workflows/sync-skill.yml` triggers on main push when `skills/**/SKILL.md` or `agents/cavecrew-*.md` changes.
123 
124What it does:
1251. Copies `skills/caveman/SKILL.md` and `skills/cavecrew/SKILL.md` into their `plugins/caveman/skills/<name>/` mirrors so the Claude Code plugin loader sees the latest behavior.
1262. Copies `skills/caveman-compress/SKILL.md` and its `scripts/` into `plugins/caveman/skills/caveman-compress/`.
1273. Copies `agents/cavecrew-*.md` into `plugins/caveman/agents/`.
1284. Rebuilds `dist/caveman.skill` (ZIP of `skills/caveman/`) for the release artifact.
1295. Commits and pushes with `[skip ci]` to avoid loops.
130 
131CI bot commits as `github-actions[bot]`. After PR merge, wait for workflow before declaring release complete.
132 
133The old steps that mirrored SKILL.md and rules into root dotdirs (`.cursor/`, `.windsurf/`, `.clinerules/`, `.github/copilot-instructions.md`) are gone — those mirrors no longer exist. The old `caveman-compress/` → `skills/compress/` rename-on-sync is also gone now that compress lives at `skills/caveman-compress/`.
134 
135---
136 
137## Hook system (Claude Code)
138 
139Three hooks in `src/hooks/` plus a `caveman-config.js` shared module and a `package.json` CommonJS marker. Communicate via flag file at `$CLAUDE_CONFIG_DIR/.caveman-active` (falls back to `~/.claude/.caveman-active`).
140 
141```
142SessionStart hook ──writes "full"──▶ $CLAUDE_CONFIG_DIR/.caveman-active ◀──writes mode── UserPromptSubmit hook
143                                                       │
144                                                    reads
145                                                       ▼
146                                              caveman-statusline.sh
147                                            [CAVEMAN] / [CAVEMAN:ULTRA] / ...
148```
149 
150`src/hooks/package.json` pins the directory to `{"type": "commonjs"}` so the `.js` hooks resolve as CJS even when an ancestor `package.json` (e.g. `~/.claude/package.json` from another plugin) declares `"type": "module"`. Without this, `require()` blows up with `ReferenceError: require is not defined in ES module scope`.
151 
152All hooks honor `CLAUDE_CONFIG_DIR` for non-default Claude Code config locations.
153 
154### `src/hooks/caveman-config.js` — shared module
155 
156Exports:
157- `getDefaultMode()` — resolves default mode in order: `CAVEMAN_DEFAULT_MODE` env var → repo-local config (`<cwd>/.caveman/config.json` or `<cwd>/.caveman.json`, walking up to the filesystem root) → user config (`$XDG_CONFIG_HOME/caveman/config.json` / `~/.config/caveman/config.json` / `%APPDATA%\caveman\config.json`) → `'full'`. The env var short-circuits before any cwd walk. Repo-local config lets a team check in a per-project default without polluting every contributor's env or user config.
158- `findRepoConfigPath(start)` — walks up from `start` (default `process.cwd()`) looking for the first `.caveman/config.json` or `.caveman.json`. Bounded to 64 ancestors. Refuses symlinked files (symmetric with `safeWriteFlag` / `readFlag`).
159- `safeWriteFlag(flagPath, content)` — symlink-safe flag write. Refuses if flag target or its immediate parent is a symlink. Opens with `O_NOFOLLOW` where supported. Atomic temp + rename. Creates with `0600`. Protects against local attackers replacing the predictable flag path with a symlink to clobber files writable by the user. Used by both write hooks. Silent-fails on all filesystem errors.
160 
161### `src/hooks/caveman-activate.js` — SessionStart hook
162 
163Runs once per Claude Code session start. Three things:
1641. Writes the active mode to `$CLAUDE_CONFIG_DIR/.caveman-active` via `safeWriteFlag` (creates if missing). Branches on the hook payload's `source` field (#691): `startup` resets to the configured default; `resume`/`clear`/`compact` re-fires preserve a valid existing flag so mid-session `/caveman <level>` switches survive.
1652. Emits caveman ruleset as hidden stdout — Claude Code injects SessionStart hook stdout as system context, invisible to user
1663. Checks `settings.json` for statusline config; if missing, appends nudge to offer setup — once per install, gated by a `.caveman-nudge-shown` marker file (#661)
167 
168Silent-fails on all filesystem errors — never blocks session start.
169 
170### `src/hooks/caveman-mode-tracker.js` — UserPromptSubmit hook
171 
172Reads JSON from stdin. Three responsibilities:
173 
174**1. Slash-command activation.** If prompt starts with `/caveman`, writes mode to flag file via `safeWriteFlag`:
175- `/caveman` → configured default (see `caveman-config.js`, defaults to `full`)
176- `/caveman lite` → `lite`
177- `/caveman ultra` → `ultra`
178- `/caveman wenyan` or `/caveman wenyan-full` → `wenyan` (alias) / `wenyan-full`
179- `/caveman wenyan-lite` → `wenyan-lite`
180- `/caveman wenyan-ultra` → `wenyan-ultra`
181- `/caveman-commit` → `commit`
182- `/caveman-review` → `review`
183- `/caveman-compress` → `compress`
184 
185**2. Natural-language activation/deactivation.** Matches phrases like "activate caveman", "turn on caveman mode", "talk like caveman" and writes the configured default mode. Matches "stop caveman", "disable caveman", "normal mode", "deactivate caveman" etc. and deletes the flag file. README promises these triggers, the hook enforces them.
186 
187**3. Per-turn reinforcement.** When flag is set to a non-independent mode (i.e. not `commit`/`review`/`compress`), emits a small `hookSpecificOutput` JSON reminder so the model keeps caveman style after other plugins inject competing instructions mid-conversation. The full ruleset still comes from SessionStart — this is just an attention anchor.
188 
189### `src/hooks/caveman-statusline.sh` — Statusline badge
190 
191Reads flag file at `$CLAUDE_CONFIG_DIR/.caveman-active`. Outputs colored badge string for Claude Code statusline:
192- `full` or empty → `[CAVEMAN]` (orange)
193- anything else → `[CAVEMAN:<MODE_UPPERCASED>]` (orange)
194 
195Then appends the lifetime-savings suffix (`⛏ 12.4k`) read from `$CLAUDE_CONFIG_DIR/.caveman-statusline-suffix` — written by `caveman-stats.js` on every `/caveman-stats` run. **Default on**; users opt out with `CAVEMAN_STATUSLINE_SAVINGS=0`. The suffix file is absent until `/caveman-stats` runs at least once, so fresh installs render no fake number.
196 
197Configured in `settings.json` under `statusLine.command`. PowerShell counterpart at `src/hooks/caveman-statusline.ps1` for Windows. Both scripts symlink-refuse and whitelist-validate the flag/suffix file contents — never echo arbitrary bytes.
198 
199### Hook installation
200 
201**Plugin install** — hooks wired automatically by plugin system.
202 
203**Standalone install** — `cli/install.js` (the unified Node installer) copies hook files into `$CLAUDE_CONFIG_DIR/hooks/` and merges SessionStart + UserPromptSubmit + statusline into `settings.json`. Uses the JSONC-tolerant helpers in `cli/lib/settings.js` so a commented `settings.json` no longer crashes the merge. Defensive `validateHookFields` runs before every write to prevent a single malformed hook from poisoning the entire file (Claude Code Zod silently discards the whole `settings.json` on schema mismatch).
204 
205The `install.sh` / `install.ps1` shims at the repo root delegate to `cli/install.js` via `node` (local clone) or `npx -y github:JuliusBrussee/caveman` (curl|bash). No legacy fallback path remains — earlier `install.sh.legacy` / `install.ps1.legacy` files were removed.
206 
207**Uninstall** — `npx -y github:JuliusBrussee/caveman -- --uninstall` (or `node cli/install.js --uninstall` from a clone). Strips caveman hook entries from `settings.json` via substring marker `caveman`, deletes hook files, and removes the Claude plugin / Gemini extension. Also removes state files from `$CLAUDE_CONFIG_DIR` (`.caveman-active`, `.caveman-active.prev`, `.caveman-mode-log.jsonl`, `.caveman-statusline-suffix`, `.caveman-nudge-shown`); keeps `.caveman-history.jsonl` (lifetime savings data) with a printed note (#635). Skill installs done via `npx skills add` must be removed via the IDE's skill manager (we don't track them).
208 
209---
210 
211## Skill system
212 
213Skills = Markdown files with YAML frontmatter consumed by Claude Code's skill/plugin system and by `npx skills` for other agents.
214 
215Each skill has a human-facing `README.md` alongside the LLM-facing `SKILL.md`. The README explains what the skill does for users browsing GitHub; the SKILL.md is the prompt body the agent loads. Don't merge them — different audiences, different formats.
216 
217### Intensity levels
218 
219Defined in `skills/caveman/SKILL.md`. Six levels: `lite`, `full` (default), `ultra`, `wenyan-lite`, `wenyan-full`, `wenyan-ultra`. Persists until changed or session ends.
220 
221### Auto-clarity rule
222 
223Caveman drops to normal prose for: security warnings, irreversible action confirmations, multi-step sequences where fragment ambiguity risks misread, user confused or repeating question. Resumes after. Defined in skill — preserve in any SKILL.md edit.
224 
225### caveman-compress
226 
227Sub-skill in `skills/caveman-compress/SKILL.md`. Takes file path, compresses prose to caveman style, writes to original path, saves backup at `<filename>.original.md`. Validates headings, code blocks, URLs, file paths, commands preserved. Retries up to 2 times on failure with targeted patches only. Requires Python 3.10+.
228 
229The slash command is `/caveman-compress` everywhere — same name in plugin and standalone install. CI no longer renames the directory on sync (the old `caveman-compress/` → `skills/compress/` sed rename is gone now that the source lives at `skills/caveman-compress/`).
230 
231### caveman-commit / caveman-review
232 
233Independent skills in `skills/caveman-commit/SKILL.md` and `skills/caveman-review/SKILL.md`. Both have own `description` and `name` frontmatter so they load independently. caveman-commit: Conventional Commits, ≤50 char subject. caveman-review: one-line comments in `L<line>: <severity> <problem>. <fix>.` format.
234 
235---
236 
237## Agent distribution
238 
239How caveman reaches each agent type:
240 
241| Agent | Mechanism | Auto-activates? |
242|-------|-----------|----------------|
243| Claude Code | Plugin (hooks + skills) or standalone hooks | Yes — SessionStart hook injects rules |
244| Codex | Plugin in `plugins/caveman/` plus repo `.codex/hooks.json` and `.codex/config.toml` | Yes on macOS/Linux — SessionStart hook |
245| Gemini CLI | Extension with `GEMINI.md` context file | Yes — context file loads every session |
246| opencode | Native plugin (`src/plugins/opencode/`) copied into `~/.config/opencode/plugins/caveman/` + `AGENTS.md` ruleset + skills/agents/commands directories. Plugin uses `session.created` and `tui.prompt.append` lifecycle hooks. No statusline (opencode TUI exposes no plugin-writable badge). | Yes — `session.created` writes flag, `AGENTS.md` carries always-on ruleset |
247| OpenClaw | Workspace skill at `~/.openclaw/workspace/skills/caveman/SKILL.md` (frontmatter merged with `version` + `always: true`) plus a marker-fenced bootstrap block in `~/.openclaw/workspace/SOUL.md`. Both writes go through `cli/lib/openclaw.js`; workspace path is overridable via `OPENCLAW_WORKSPACE`. | Yes — SOUL.md is auto-injected each turn under "Project Context" (subject to OpenClaw's 12K-per-file / 60K-total bootstrap caps) |
248| Cursor | `npx skills add ... -a cursor` (default via `--only cursor`) writes the upstream skill profile; per-repo `.cursor/rules/caveman.mdc` via `--with-init` (calls `src/tools/caveman-init.js`) | Yes — always-on rule |
249| Windsurf | `npx skills add ... -a windsurf` (default via `--only windsurf`); per-repo `.windsurf/rules/caveman.md` via `--with-init` | Yes — always-on rule |
250| Cline | `npx skills add ... -a cline` (default via `--only cline`); per-repo `.clinerules/caveman.md` via `--with-init` | Yes — Cline auto-discovers `.clinerules/` |
251| Copilot | `npx skills add ... -a github-copilot` (soft probe — pass `--only copilot`); per-repo `.github/copilot-instructions.md` + `AGENTS.md` via `--with-init` | Yes — repo-wide instructions |
252| Others (Junie, Trae, Warp, Tabnine, Mistral, Qwen, Devin, Droid, ForgeCode, Bob, Crush, iFlow, OpenHands, Qoder, Rovo Dev, Replit, Antigravity, …) | `npx skills add JuliusBrussee/caveman -a <profile>` | No — user must say `/caveman` each session |
253 
254opencode reaches Tier 1 minus the statusline (opencode's TUI has no plugin-writable badge). Mode flag lives at `~/.config/opencode/.caveman-active` for any external tooling that wants to surface it.
255 
256For agents without hook systems, the always-on snippet lives in `INSTALL.md`'s "Want it always on?" section — keep current with `src/rules/caveman-activate.md`.
257 
258**Adding a new agent.** Edit the `PROVIDERS` array in `cli/install.js` — single source of truth, no more bash/PS1 dual-source drift. Each entry has `id`, `label`, `mech`, `detect` (clause spec like `command:foo||dir:$HOME/x`), optional `profile` (vercel-labs/skills slug), optional `soft: true` (config-dir-only detection).
259 
2601. The profile slug must exist in upstream [vercel-labs/skills](https://github.com/vercel-labs/skills). Verify against the README before merging — wrong slugs cause `npx skills add` to fail at runtime, not at install-script load.
2612. Run `node cli/install.js --list` to confirm the new row renders correctly.
2623. Soft probes (config-dir-only) are fine but tag them with `soft: true`. They render with `(soft)` in `--list` so users know detection is best-effort.
263 
264---
265 
266## Evals
267 
268`evals/` has three-arm harness:
269- `__baseline__` — no system prompt
270- `__terse__` — `Answer concisely.`
271- `<skill>` — `Answer concisely.\n\n{SKILL.md}`
272 
273Honest delta = **skill vs terse**, not skill vs baseline. Baseline comparison conflates skill with generic terseness — that cheating. Harness designed to prevent this.
274 
275`llm_run.py` calls `claude -p --system-prompt ...` per (prompt, arm), saves to `evals/snapshots/results.json`. `measure.py` reads snapshot offline with tiktoken (OpenAI BPE — approximates Claude tokenizer, ratios meaningful, absolute numbers approximate).
276 
277Add skill: drop `skills/<name>/SKILL.md`. Harness auto-discovers. Add prompt: append line to `evals/prompts/en.txt`.
278 
279Snapshots committed to git. CI reads without API calls. Only regenerate when SKILL.md or prompts change.
280 
281---
282 
283## Benchmarks
284 
285`benchmarks/` runs real prompts through Claude API (not Claude Code CLI), records raw token counts. Results committed as JSON in `benchmarks/results/`. Benchmark table in README generated from results — update when regenerating.
286 
287To reproduce: `uv run python benchmarks/run.py` (needs `ANTHROPIC_API_KEY` in `.env.local`).
288 
289---
290 
291## Key rules for agents working here
292 
293- Edit `skills/<name>/SKILL.md` for behavior changes. Never edit synced copies under `plugins/caveman/skills/`.
294- Edit `src/rules/caveman-activate.md` for auto-activation rule changes. Never edit any per-agent rule copy a user has on their machine.
295- Edit `src/rules/caveman-openclaw-bootstrap.md` for the OpenClaw SOUL.md bootstrap snippet. Keep the `<!-- caveman-begin -->` / `<!-- caveman-end -->` markers and the `Respond terse like smart caveman` sentinel — `cli/lib/openclaw.js` keys idempotency off both. If you change the embedded fallback in `cli/lib/openclaw.js`, keep it byte-equivalent to the file.
296- Per-skill human docs live in `skills/<name>/README.md`. The LLM-facing body is in `SKILL.md`. Don't merge them — different audiences.
297- Build artifacts go in `dist/`. Never check files into `dist/` manually — CI rebuilds them on push, and `dist/` is gitignored.
298- README most important file for user-facing impact. Optimize for non-technical readers. Preserve caveman voice.
299- `INSTALL.md` is the per-agent install reference. Keep the install table in `README.md` short and link out to `INSTALL.md` for the full matrix.
300- Benchmark and eval numbers must be real. Never fabricate or estimate.
301- CI workflow commits back to main after merge. Account for when checking branch state.
302- Hook files must silent-fail on all filesystem errors. Never let hook crash block session start.
303- Any new flag file write must go through `safeWriteFlag()` in `caveman-config.js`. Direct `fs.writeFileSync` on predictable user-owned paths reopens the symlink-clobber attack surface.
304- Hooks must respect `CLAUDE_CONFIG_DIR` env var, not hardcode `~/.claude`. Same for `cli/install.js` / statusline scripts.
305- `cli/install.js` is the only installer source. `install.sh` / `install.ps1` at repo root are 30-line shims that delegate to it. Never re-add per-OS install logic to the shims — that's how we got the Windows quoting bug (#249).
306- Any settings.json read in installer or hooks must go through `cli/lib/settings.js` `readSettings()` so JSONC comments don't crash the merge. Any settings.json write must run through `validateHookFields()` first.
307