# Bash Script Installer for Zsh The "Bash Script Installer" simplifies the setup of Zsh. ## Zsh ### Ubuntu ```zsh bash -c "$(curl -fsSL https://tinyurl.com/y6fhb594)" ``` ### WSL ```zsh bash -c "$(curl -fsSL https://tinyurl.com/ycyct6wp)" ``` ### MacOS ```zsh bash -c "$(curl -fsSL https://tinyurl.com/yptybtpa)" ``` ### Configuration ```zsh bash -c "$(curl -fsSL https://tinyurl.com/y8vf753j)" ``` ### Shell framework and prompt The installers use Oh My Zsh for framework features and plugins, with the OMZ theme disabled: ```zsh ZSH_THEME="" source "$ZSH/oh-my-zsh.sh" eval "$(starship init zsh)" ``` Starship owns the prompt and loads after Oh My Zsh. The managed Starship config is installed at `~/.config/starship.toml`. A Nerd Font is required for the Powerline symbols. ## AI commits The managed `aicommit` function can use Claude, Copilot, or OpenCode to create a commit from staged changes: ```zsh aicommit opencode ``` The installers place the shared commit instructions at `~/.config/ai-commit-prompt.txt` and the required OpenCode agent at `~/.config/opencode/agents/ai-commit.md`. The agent is hidden from the interactive agent selector but remains available to `aicommit opencode`. It may inspect Git status, the cached diff, and recent history, and may execute `git commit`; all unrelated tools and shell commands are denied. **Changes affecting more than 30 staged paths automatically use per-module commits.** They must produce multiple commits, rather than a single combined commit message. The count uses the full staged set before diff filtering or truncation, including lock/minified files and both paths of a rename. Exactly 30 paths retains the single-commit default. Pass `--modules` to request splitting at any size: ```zsh aicommit --modules opencode aicommit --modules --yes claude ``` Automatic splitting uses the same planning and confirmation flow as `--modules`; `--yes` skips confirmation in either case (for example, `aicommit --yes opencode`). Plans that collapse a change above the threshold into one group are rejected before committing, with the index untouched. Grouping follows cohesive modules or submodules; 30 is the automatic-splitting threshold, not a hard size limit per planned module. The engine first plans the split using `~/.config/ai-commit-split-prompt.txt` and replies with a JSON plan that maps every staged path to a `type(scope)` group, ordered so dependencies land first. The function validates the plan against the staged paths, ignores paths that are not staged, appends any unplanned paths as a final commit, and shows the plan for confirmation (`--yes` skips it). It then stages each group from a snapshot of the original index, so partially staged files keep only their staged hunks and unstaged or untracked work is never touched, and asks the engine to commit that group with the normal commit instructions. If a group produces no commit, or a commit includes files outside its group, the run stops and restages the remaining groups. With OpenCode, planning uses the hidden `ai-commit-plan` agent at `~/.config/opencode/agents/ai-commit-plan.md`, which is denied every tool. Python 3 is required to parse the plan. ## Plan-build workflow The local `plan_build` function is a thin cached launcher for the canonical public [plan-build gist](https://opengist.resetrix.work/weehong/plan-build). When curl is available, each invocation stages `plan_build.zsh`, `SKILL.md`, and `ARCHITECT.md`, requires all three to be non-empty, and syntax-checks the executable. It then moves the complete bundle into a unique immutable release directory and atomically switches one `current` symlink. Refresh activation is serialized with a kernel-backed Zsh file lock so concurrent callers cannot mix releases and crashes cannot leave a stale lock. If refresh fails, the launcher preserves and uses the last complete, validated release and repairs its skill links. Zsh is always required; curl is required only when no valid cached release exists. ```zsh plan_build plan_build --prompt plan_build --writing-plan plan_build --prompt --brainstorm plan_build --yolo --writing-plan plan_build --architect plan_build --architect --new plan_build --architect --yolo ``` The canonical executable defines the current command behavior and options. Each invocation binds the executable and both skills to its selected immutable release. The launcher also points the installed Claude Code skill paths through the active `current` release: - `SKILL.md` → `~/.claude/skills/plan-build/SKILL.md` - `ARCHITECT.md` → `~/.claude/skills/plan-build-architect/SKILL.md` The launcher defaults to `https://opengist.resetrix.work/weehong/plan-build/raw/HEAD` and `${XDG_CACHE_HOME:-$HOME/.cache}/plan-build`. Override these with `PLAN_BUILD_BASE_URL` and `PLAN_BUILD_CACHE_DIR`. Installed skill paths remain fixed at `~/.claude/skills`; cached invocations receive immutable release paths through environment variables. The downloaded script is remote executable code: syntax validation catches malformed Zsh but is not a security sandbox, so only point `PLAN_BUILD_BASE_URL` at a source you trust. Remote code is executed in a child Zsh process and is never sourced into the caller shell. Invoking `plan_build` once installs or repairs both skill links. To install standalone files without the cached launcher, follow the installation instructions in the canonical gist instead.

--- description: Plans how to split the staged diff supplied in the prompt into per-module commits. mode: primary hidden: true permission: "*": deny --- Follow the supplied planning instructions and staged file list. Reply with only the JSON plan. Do not run commands, modify files, or change the staged set. Changes affecting more than 30 staged paths must be modularized into at least two non-empty commits; never return a single combined group. Count the full staged file list, including lock/minified files and both paths of a rename. Split a large module into cohesive, independently reviewable submodules. At 30 paths or fewer, a single group is allowed if the changes form one module. This threshold is not a hard file limit per planned module.

# Git Commit Message Generator Prompt ## Role You are an Automated Git Commit Engine. ## Objective Analyze staged changes (`git diff --cached`), identify the **primary logical change**, and generate a commit message that **strictly adheres** to Conventional Commits v1.0.0. --- ## Input The user will provide the output of: ```bash git diff --cached ``` --- ## Strict Rules (Zero Tolerance) ### Mandatory Module Splitting Changes affecting more than 30 staged paths MUST be modularized into multiple commits. Do not generate or execute a single combined commit for them. Count the full staged set, including lock/minified files omitted from the diff preview and both paths of a rename, not just the visible diff. If this input has more than 30 staged paths and is not explicitly identified as one module in a multi-commit plan, stop and request `aicommit --modules`. This refusal takes precedence over the output and execution instructions below. When the input says "commit X of Y, split by module" with Y greater than 1, generate and execute only that module's commit. The threshold applies to the original staged change, not a hard file limit per planned module. ### 1. Header Format **Pattern:** `type(scope): subject` | Element | Rule | | :--- | :--- | | **Type** | ONLY use: `feat`, `fix`, `docs`, `style`, `refactor`, `perf`, `test`, `build`, `ci`, `chore` | | **Scope** | Optional, in parentheses, entirely lowercase (e.g., `docker`, `api`, `auth`) | | **Subject** | Lowercase, imperative mood ("add" not "added"), NO period at the end | | **Length** | Header MUST be ≤ 50 characters | ### 2. Body Format | Element | Rule | | :--- | :--- | | **Spacing** | Exactly ONE blank line between the header and the body | | **Bullet Points**| Use hyphens (`-`) only | | **Line Length** | Each line MUST be ≤ 72 characters | | **Case** | Lowercase throughout | ### 3. Content Rules * **NO** "Merge branch" or auto-generated text. * **NO** Markdown formatting (bold/italic/code blocks) inside the commit message string itself. * **NO** periods at the end of the subject line. * **NO** "Co-authored-by" trailers or any co-author attributions. --- ## Execution Steps ### Step 1: Atomic Scan Identify the **largest cluster of related changes** (the "Primary Scope"). **Priority Order for Type Selection:** 1. `ci` - Pipeline, CI/CD configuration changes 2. `feat` - New features 3. `fix` - Bug fixes 4. `refactor` - Code restructuring 5. `build` - Build system, dependencies 6. `test` - Test additions/modifications 7. `chore` - Maintenance tasks 8. `docs` - Documentation only (use ONLY if no code changes exist) ### Step 2: Classify Changes Categorize all changed files: | Category | Influences Subject? | Include in Body? | | :--- | :---: | :---: | | **Primary Scope** | ✅ Yes | ✅ Yes | | **Supporting Changes** | ❌ No | ✅ Yes | | **Documentation** | ❌ No | ✅ Yes (as a single line) | > **Key Rule:** Documentation changes (`docs/`) should be mentioned in the body but MUST NOT influence the subject type unless documentation is the *only* change in the commit. ### Step 3: Draft Commit Message Structure the message exactly like this: ```text type(scope): subject (≤50 chars, imperative, lowercase, no period) - primary change detail 1 - primary change detail 2 - supporting change detail - add/update project documentation ``` --- ## Output Format & Execution Authorization You must output **ONLY** a single bash code block containing the complete git command. Do not provide any conversational filler, explanations, or greeting. Once you have generated the `git commit` bash command, **you are explicitly authorized to execute it directly in the terminal** (if your environment supports agentic execution). * Do not ask for further user confirmation. * Execute sequentially. * If the command fails, output the exact error message and halt execution. **Format Template:** ```bash git commit -m "type(scope): subject - body detail 1 - body detail 2 - body detail 3" ``` --- ## Examples ### Example 1: CI/CD with Documentation * **Changes:** `.gitlab-ci.yaml` (new), `Dockerfile` (new), `docs/deployment.md` (new) * **Output:** ```bash git commit -m "ci: add gitlab pipeline and docker config - add gitlab-ci.yaml with build, test, publish stages - add Dockerfile for container deployment - add deployment documentation" ``` ### Example 2: Feature with Config Changes * **Changes:** `src/auth/login.java` (new), `pom.xml` (modified), `docs/api.md` (new) * **Output:** ```bash git commit -m "feat(auth): add user login endpoint - implement login service with jwt tokens - add spring-security dependency to pom.xml - add api documentation" ``` ### Example 3: Documentation Only * **Changes:** `docs/setup.md` (new), `docs/testing.md` (modified), `README.md` (modified) * **Output:** ```bash git commit -m "docs: add setup guide and update testing docs - add initial setup guide - update testing documentation with examples - update readme with quick start section" ``` --- ## Quick Reference | Type | When to Use | | :--- | :--- | | `feat` | New feature for the user | | `fix` | Bug fix for the user | | `docs` | Documentation only changes | | `style` | Formatting, missing semicolons, etc. (no code change) | | `refactor` | Code change that neither fixes a bug nor adds a feature | | `perf` | Code change that improves performance | | `test` | Adding or refactoring tests | | `build` | Build system or external dependencies | | `ci` | CI configuration files and scripts | | `chore` | Maintenance tasks, no production code changes |

# Git Commit Split Planner Prompt ## Role You are an Automated Git Commit Planner. ## Objective Split the staged changes into a sequence of small, atomic commits, one per **module**. You only decide the grouping and order. You do NOT write commit messages and you do NOT run any commands. --- ## Input 1. The staged file list (`git diff --cached --name-status -M`) 2. A diffstat (`git diff --cached --numstat -M`: added, deleted, path) 3. The staged diff (lock files and minified files are omitted; it may be truncated) --- ## What Is a Module A module is one cohesive, independently reviewable unit of change that maps to a single Conventional Commit `type(scope)`. Typical modules: | Module | Example scope | | :--- | :--- | | A feature or domain area (pages, routes, services, models) | `organisation`, `auth` | | Shared UI components / design system | `ui` | | Translations / localisation | `i18n` | | End-to-end or integration tests not tied to one module | `e2e` | | Storybook stories | `storybook` | | Documentation | `docs` | | Build, dependencies, tooling config | `deps`, `eslint` | | CI/CD pipelines | `ci` | --- ## Strict Rules (Zero Tolerance) 1. Every path in the staged file list MUST appear in exactly one commit. Copy paths exactly as listed. Never invent, rename, or omit a path. 2. A rename line (`R<score> old new`) lists two paths. Put BOTH in the same commit. 3. Never split a single file across commits. 4. Lock files go in the same commit as the manifest that changed them (e.g. `package.json` with `package-lock.json`). 5. Unit tests that sit next to, or only exercise, one module go in that module's commit. E2E tests, storybook stories and docs get their own commits unless they are the only change. 6. Each commit should have ONE type and ONE scope. If a group would need two types (e.g. `feat` and `docs`), split it. 7. Do not over-split. A module normally spans several files; do not create one commit per file unless each file really is its own module. Changes affecting more than 30 staged paths MUST be modularized into at least two non-empty commits; never return a single combined commit for them. Count all staged paths, including lock/minified files and both paths of a rename. Split a large module into cohesive, independently reviewable submodules. At 30 paths or fewer, a single commit is allowed if all changes form one module. This threshold is not a hard file limit per planned module. 8. Order commits so dependencies land first: build/deps -> shared code (ui, i18n, lib, types) -> features that use them -> tests -> stories/docs. --- ## Output Format Output ONLY one JSON object. No prose, no explanation, no markdown fences. ```json {"commits":[{"type":"feat","scope":"i18n","files":["src/locales/en/organisation.json"]}]} ``` * `type`: one of `feat`, `fix`, `docs`, `style`, `refactor`, `perf`, `test`, `build`, `ci`, `chore` * `scope`: lowercase scope, or an empty string if none fits * `files`: the exact paths from the staged file list --- ## Example * **Staged files:** `src/locales/en/organisation.json`, `src/locales/en/settings.json`, `src/components/ui/data-table.tsx`, `src/components/ui/stat-card.tsx`, `src/app/organisation/page.tsx`, `src/app/organisation/list/page.tsx`, `src/components/ui/data-table.stories.tsx`, `e2e/home.spec.ts` * **Output:** ```json {"commits":[ {"type":"feat","scope":"i18n","files":["src/locales/en/organisation.json","src/locales/en/settings.json"]}, {"type":"feat","scope":"ui","files":["src/components/ui/data-table.tsx","src/components/ui/stat-card.tsx"]}, {"type":"feat","scope":"organisation","files":["src/app/organisation/page.tsx","src/app/organisation/list/page.tsx"]}, {"type":"docs","scope":"storybook","files":["src/components/ui/data-table.stories.tsx"]}, {"type":"test","scope":"e2e","files":["e2e/home.spec.ts"]} ]} ```

--- description: Creates one conventional commit from the staged diff supplied in the prompt. mode: primary hidden: true permission: "*": deny bash: "*": deny "git status *": allow "git diff --cached*": allow "git log *": allow "git commit *": allow --- Follow the supplied commit instructions and staged diff. Changes affecting more than 30 staged paths must be modularized into multiple commits; never emit or execute a single combined commit for them. Use the supplied full staged path count, or inspect `git diff --cached --name-only --no-renames` if it is absent, including files omitted from the diff preview. Unless the input explicitly identifies this as one module in a multi-commit plan ("commit X of Y, split by module", Y > 1), stop and request `aicommit --modules` when the count exceeds 30. This refusal takes precedence over commit execution. For an eligible staged change or an individual module from that plan, generate a concise conventional commit message, then execute exactly one `git commit` with that message. You may inspect Git status, the cached diff, and recent log. Do not modify files or change the staged set.

# ============================================================================= # OS-SPECIFIC ALIASES # ============================================================================= if [[ "$OSTYPE" == "linux-gnu"* ]] && command -v apt-get >/dev/null 2>&1; then # Linux Only System update & cleanup alias uu='sudo apt-get update && \ sudo apt-get upgrade -y && \ sudo apt-get full-upgrade -y && \ sudo apt-get autoremove -y && \ sudo apt-get autoclean -y && \ sudo apt-get clean' elif [[ "$OSTYPE" == "darwin"* ]]; then # macOS Only alias flushdns='sudo dscacheutil -flushcache; sudo killall -HUP mDNSResponder' fi # ============================================================================= # UNIVERSAL GIT ALIASES (Fixed: Removed broken "$@" from aliases) # ============================================================================= alias gph='git push' alias gco='git checkout' alias gbh='git branch' alias gmt='git commit' alias gpl='git pull' alias grb='git rebase' alias grt='git reset' alias gst='git status' alias gsts='git stash push --staged' # Keep staged changes in the index; the stash also records their snapshot. alias gstu='git stash push --keep-index --include-untracked' alias gsta='git stash push --include-untracked' alias grmrf='git checkout -- . && git clean -fd' # Create a feature worktree from origin/integration: gwa pill gwa() { if [ "$#" -ne 1 ] || [ -z "$1" ]; then printf 'Usage: gwa <feature-name>\n' >&2 return 1 fi git worktree add "../worktree/$1" -b "feature/$1" origin/integration } # List loaded aliases, optionally filtered by a case-insensitive keyword. ah() { if [ "$#" -eq 0 ]; then alias else alias | command grep -iF -- "$*" fi }

#!/bin/bash set -euo pipefail # Configuration GIST_RAW_BASE="https://opengist.resetrix.work/weehong/f0d940c3c1214bf5b7996195199fdc09/raw/HEAD" CONFIG_FILES=( ".alias" ".func" ".pathrc" ".sourcerc" ".vimrc" ".zshrc" ".config/starship.toml" ".config/ai-commit-prompt.txt" ".config/ai-commit-split-prompt.txt" ".config/opencode/agents/ai-commit.md" ".config/opencode/agents/ai-commit-plan.md" ) echo "Starting configuration download..." download_failed=0 for f in "${CONFIG_FILES[@]}"; do # Remove the leading dot for the URL path remote_name="${f#.}" [[ "$f" == ".config/starship.toml" ]] && remote_name="starship.toml" [[ "$f" == ".config/ai-commit-prompt.txt" ]] && remote_name="ai-commit-prompt.txt" [[ "$f" == ".config/ai-commit-split-prompt.txt" ]] && remote_name="ai-commit-split-prompt.txt" [[ "$f" == ".config/opencode/agents/ai-commit.md" ]] && remote_name="ai-commit.md" [[ "$f" == ".config/opencode/agents/ai-commit-plan.md" ]] && remote_name="ai-commit-plan.md" url="$GIST_RAW_BASE/$remote_name" target="$HOME/$f" tmp="${target}.tmp.$$" echo "Downloading $f..." mkdir -p "$(dirname "$target")" # Use -f to fail silently on server errors, -s for silent, -L to follow redirects if curl -fsSL "$url" -o "$tmp" && mv "$tmp" "$target"; then echo "Successfully updated $target" else rm -f "$tmp" echo "Error: Failed to download $f from $url" >&2 download_failed=1 fi done if [ "$download_failed" -ne 0 ]; then echo "Configuration download completed with errors." >&2 exit 1 fi echo "Done! All configuration files have been replaced."

# ============================================================================= # CUSTOM FUNCTIONS # ============================================================================= # ----------------------------------------------------------------------------- # Function: funcs (List available .func functions) # ----------------------------------------------------------------------------- funcs() { local func_file="${1:-$HOME/.func}" if [ ! -f "$func_file" ]; then printf "Error: Function file '%s' not found.\n" "$func_file" >&2 return 1 fi printf "Available functions in %s:\n" "$func_file" awk ' /^# Function: / { line = $0 sub(/^# Function: /, "", line) name = line sub(/ .*/, "", name) desc = "" if (line ~ / \(.+\)$/) { desc = line sub(/^[^ ]+ \(/, "", desc) sub(/\)$/, "", desc) } if (desc != "") { printf " %-22s %s\n", name, desc } else { printf " %s\n", name } } ' "$func_file" } # ----------------------------------------------------------------------------- # Function: clip (Cross-platform clipboard) # ----------------------------------------------------------------------------- clip() { local cmd local args=() if command -v pbcopy >/dev/null 2>&1; then cmd="pbcopy" # macOS elif grep -qi "microsoft" /proc/version 2>/dev/null && command -v clip.exe >/dev/null 2>&1; then cmd="clip.exe" # WSL elif [ "$XDG_SESSION_TYPE" = "wayland" ] && command -v wl-copy >/dev/null 2>&1; then cmd="wl-copy" # Linux Wayland elif command -v xclip >/dev/null 2>&1; then cmd="xclip" # Linux X11 (Fallback 1) args=("-selection" "clipboard") elif command -v xsel >/dev/null 2>&1; then cmd="xsel" # Linux X11 (Fallback 2) args=("--clipboard" "--input") else printf "Error: No supported clipboard utility found.\n" >&2 return 1 fi if [ $# -gt 0 ]; then if [ -f "$1" ]; then "$cmd" "${args[@]}" < "$1" echo "Copied contents of '$1' to clipboard." else printf "Error: File '%s' not found.\n" "$1" >&2 return 1 fi else "$cmd" "${args[@]}" fi } # ----------------------------------------------------------------------------- # Function: open_file (Cross-platform file/directory opener) # ----------------------------------------------------------------------------- open_file() { local target="${1:-.}" if [[ "$OSTYPE" == "darwin"* ]]; then command open "$target" elif grep -qi "microsoft" /proc/version 2>/dev/null; then if command -v wslpath >/dev/null 2>&1 && command -v explorer.exe >/dev/null 2>&1; then explorer.exe "$(wslpath -w "$target")" else printf "Error: 'wslpath' or 'explorer.exe' not found.\n" >&2 return 1 fi elif [[ "$OSTYPE" == "linux-gnu"* ]]; then if command -v xdg-open >/dev/null 2>&1; then xdg-open "$target" else printf "Error: 'xdg-open' not found.\n" >&2 return 1 fi fi } alias open='open_file' # ----------------------------------------------------------------------------- # Function: ch (Supports shell, Claude, Codex, and OpenCode) # ----------------------------------------------------------------------------- ch() { if [ "$#" -gt 1 ]; then printf "Usage: ch [--all | shell | claude | codex | opencode]\n" >&2 return 2 fi case "${1:-}" in --all) local result=0 target for target in claude codex opencode shell; do ch "$target" || result=1 done if [ -t 1 ]; then command clear fi return "$result" ;; claude) local claude_home="${CLAUDE_CONFIG_DIR:-$HOME/.claude}" if [ -d "$claude_home/projects" ]; then command rm -rf -- "$claude_home/projects"/*(DN) || return 1 fi if [ -f "$claude_home/history.jsonl" ]; then : > "$claude_home/history.jsonl" || return 1 fi echo "Claude project history/cache cleared." ;; codex) local codex_home="${CODEX_HOME:-$HOME/.codex}" command mkdir -p -- "$codex_home" || return 1 : > "$codex_home/history.jsonl" || return 1 command rm -rf -- "$codex_home/sessions" "$codex_home/archived_sessions" "$codex_home/shell_snapshots" || return 1 command mkdir -p -- "$codex_home/sessions" "$codex_home/archived_sessions" "$codex_home/shell_snapshots" || return 1 echo "Codex history cleared." ;; opencode) local opencode_state_home="${XDG_STATE_HOME:-$HOME/.local/state}/opencode" local opencode_data_home="${XDG_DATA_HOME:-$HOME/.local/share}/opencode" command mkdir -p -- "$opencode_state_home" "$opencode_data_home" || return 1 : > "$opencode_state_home/prompt-history.jsonl" || return 1 command rm -f -- \ "$opencode_data_home/opencode.db" \ "$opencode_data_home/opencode.db-shm" \ "$opencode_data_home/opencode.db-wal" || return 1 command rm -rf -- "$opencode_data_home/repos" "$opencode_data_home/log" "$opencode_data_home/snapshot" || return 1 command mkdir -p -- "$opencode_data_home/repos" "$opencode_data_home/log" "$opencode_data_home/snapshot" || return 1 echo "OpenCode history cleared." ;; ""|shell) # 1. Truncate the file if [ -n "${HISTFILE:-}" ]; then : > "$HISTFILE" || return 1 fi # 2. Start a fresh Zsh history context. HISTSIZE=0 retains one entry. builtin fc -p "${HISTFILE:-}" "$HISTSIZE" "$SAVEHIST" || return 1 echo "Shell history cleared." ;; *) printf "Usage: ch [--all | shell | claude | codex | opencode]\n" >&2 return 2 ;; esac if [ -t 1 ]; then command clear fi } # ----------------------------------------------------------------------------- # Function: require_cli (Check external CLI availability) # ----------------------------------------------------------------------------- require_cli() { local binary="$1" local label="$2" local resolved if command -v whence >/dev/null 2>&1; then if whence -p "$binary" >/dev/null 2>&1; then return 0 fi else resolved="$(command -v "$binary" 2>/dev/null)" || resolved="" if [ -n "$resolved" ] && [ -x "$resolved" ]; then return 0 fi fi printf "Error: %s ('%s') is not installed or not in your PATH.\n" "$label" "$binary" >&2 return 1 } # ----------------------------------------------------------------------------- # Function: claude (Includes --yolo and clear shortcuts) # ----------------------------------------------------------------------------- claude() { # Shortcut for clearing cache if [[ "$1" == "clear" ]]; then ch claude return $? fi # Check if binary exists require_cli claude "Claude Code CLI" || return 1 # Handle --yolo mode if [[ "$1" == "--yolo" ]]; then shift command claude --dangerously-skip-permissions "$@" return $? fi command claude "$@" } # ----------------------------------------------------------------------------- # Internal: _prompt_confirm_indexing (Confirm project indexing via the terminal) # ----------------------------------------------------------------------------- _prompt_confirm_indexing() { local project_root="$1" local reply if ! (: </dev/tty) 2>/dev/null; then printf "Notice: No interactive terminal available; enhancing without project indexing.\n" >&2 return 1 fi printf "Allow Auggie to index and use project context from '%s'? (y/N) " "$project_root" >/dev/tty if ! IFS= read -r reply </dev/tty; then printf "\nNotice: Unable to read confirmation; enhancing without project indexing.\n" >&2 return 1 fi case "$reply" in y|Y|yes|YES|Yes) return 0 ;; *) return 1 ;; esac } # ----------------------------------------------------------------------------- # Internal: _enhance_prompt (Run Auggie and write only the enhanced prompt) # ----------------------------------------------------------------------------- _enhance_prompt() { local prompt="$1" local output_file="$2" local run_dir local prompt_file local workspace local cache_dir local auth_file local log_file local parsed_file local project_root local use_project_context local script_pid local waited local timeout_seconds local monitor_status if [ -z "$prompt" ]; then printf "Error: Prompt was empty.\n" >&2 return 1 fi if [ ! -s "$HOME/.augment/session.json" ]; then printf "Error: Auggie session file not found. Run 'auggie login' first.\n" >&2 return 1 fi run_dir="$(mktemp -d "${TMPDIR:-/tmp}/auggie-enhance.XXXXXX")" || return 1 prompt_file="$run_dir/prompt.txt" workspace="$run_dir/workspace" cache_dir="$run_dir/cache" auth_file="$HOME/.augment/session.json" log_file="$run_dir/auggie.log" parsed_file="$run_dir/enhanced.txt" mkdir -p "$workspace" "$cache_dir" printf "%s\n" "$prompt" > "$prompt_file" project_root="$(git rev-parse --show-toplevel 2>/dev/null)" || project_root="$PWD" use_project_context=0 timeout_seconds=90 if _prompt_confirm_indexing "$project_root"; then use_project_context=1 timeout_seconds=300 workspace="$project_root" cache_dir="$HOME/.augment" printf "Indexing approved; enhancing with project context from '%s'.\n" "$project_root" >&2 else printf "Enhancing without project indexing.\n" >&2 fi ( if [ "$use_project_context" -eq 1 ]; then AUGGIE_PROMPT_FILE="$prompt_file" \ AUGGIE_WORKSPACE="$workspace" \ AUGGIE_CACHE_DIR="$cache_dir" \ AUGGIE_AUTH_FILE="$auth_file" \ script -q -f -O "$log_file" -c 'auggie --print --enhance-prompt --workspace-root "$AUGGIE_WORKSPACE" --augment-cache-dir "$AUGGIE_CACHE_DIR" --augment-session-json "$AUGGIE_AUTH_FILE" --allow-indexing --wait-for-indexing --dont-save-session --instruction-file "$AUGGIE_PROMPT_FILE"' </dev/null >/dev/null 2>&1 & else AUGGIE_PROMPT_FILE="$prompt_file" \ AUGGIE_WORKSPACE="$workspace" \ AUGGIE_CACHE_DIR="$cache_dir" \ AUGGIE_AUTH_FILE="$auth_file" \ script -q -f -O "$log_file" -c 'auggie --print --enhance-prompt --no-discover-workspaces --workspace-root "$AUGGIE_WORKSPACE" --augment-cache-dir "$AUGGIE_CACHE_DIR" --augment-session-json "$AUGGIE_AUTH_FILE" --dont-save-session --instruction-file "$AUGGIE_PROMPT_FILE"' </dev/null >/dev/null 2>&1 & fi script_pid=$! waited=0 while kill -0 "$script_pid" >/dev/null 2>&1; do if grep -aq "🤖" "$log_file" 2>/dev/null || grep -aq "Tool call:" "$log_file" 2>/dev/null; then kill "$script_pid" >/dev/null 2>&1 || true wait "$script_pid" >/dev/null 2>&1 || true exit 0 fi if [ "$waited" -ge "$timeout_seconds" ]; then kill "$script_pid" >/dev/null 2>&1 || true wait "$script_pid" >/dev/null 2>&1 || true exit 124 fi sleep 1 waited=$((waited + 1)) done wait "$script_pid" >/dev/null 2>&1 || true exit 0 ) monitor_status=$? if [ "$monitor_status" -eq 124 ]; then printf "Error: Timed out waiting for Auggie to enhance the prompt.\n" >&2 rm -rf "$run_dir" return 124 fi perl -ne ' # Auggie redraws streamed terminal lines with carriage returns. Keep only # the final rendered segment so stale text cannot corrupt the prompt. s/\r$//; s/.*\r//; 1 while s/[^\x08]\x08//g; s/\x08//g; s/\e\][^\a]*(?:\a|\e\\)//g; s/\e\[[0-?]*[ -\/]*[@-~]//g; next if /Script started on/ || /Script done on/; if (/^(?:✨\s*)?Enhanced prompt:\s*(.*)$/) { $capturing = 1; $output .= "$1\n" if length $1; next; } next unless $capturing; exit if /^🤖/ || /Tool call:/ || /Session terminated/; $output .= $_; END { $output =~ s/^\s*\n//; $output =~ s/\s+\z//; print "$output\n" if length $output; } ' "$log_file" > "$parsed_file" if [ ! -s "$parsed_file" ]; then printf "Error: Auggie did not return an enhanced prompt.\n" >&2 rm -rf "$run_dir" return 1 fi if ! command cp "$parsed_file" "$output_file"; then printf "Error: Unable to save the enhanced prompt.\n" >&2 rm -rf "$run_dir" return 1 fi rm -rf "$run_dir" } # ----------------------------------------------------------------------------- # Function: prompt (Enhance a prompt with Auggie and optional project context) # ----------------------------------------------------------------------------- prompt() { local prompt local line local output_file local input_keymap local input_status require_cli auggie "Auggie CLI" || return 1 require_cli script "script utility" || return 1 if [ $# -gt 0 ]; then prompt="$*" elif [ ! -t 0 ]; then prompt="$(cat)" elif [ -n "${ZSH_VERSION:-}" ]; then # Keep all lines in one ZLE buffer so previous lines remain editable. input_keymap="auggie-prompt-$$-$RANDOM" bindkey -N "$input_keymap" emacs || return 1 bindkey -M "$input_keymap" '^J' self-insert bindkey -M "$input_keymap" '^M' self-insert-unmeta bindkey -M "$input_keymap" '^[[13;2u' accept-line bindkey -M "$input_keymap" '^[[27;2;13~' accept-line bindkey -M "$input_keymap" '^D' accept-line printf 'Enter your prompt: Enter adds a line; Shift+Enter (configured terminal) or Ctrl+D submits; Ctrl+C cancels.\n' prompt="" vared -M "$input_keymap" -p '> ' prompt input_status=$? bindkey -D "$input_keymap" [ "$input_status" -eq 0 ] || return "$input_status" else echo "Reading prompt... Type 'EOF' on a new line and press Enter when finished." prompt="" while IFS= read -r line; do [ "$line" = "EOF" ] && break if [ -z "$prompt" ]; then prompt="$line" else prompt="${prompt}"$'\n'"${line}" fi done fi if [ -z "$prompt" ]; then printf "Error: Prompt was empty.\n" >&2 return 1 fi output_file="$(mktemp "${TMPDIR:-/tmp}/auggie-enhanced-prompt.XXXXXX")" || return 1 if ! _enhance_prompt "$prompt" "$output_file"; then rm -f "$output_file" return 1 fi if [ -t 1 ]; then clear fi printf "Enhanced prompt:\n" command cat "$output_file" if clip < "$output_file"; then printf '\nEnhanced prompt copied to clipboard.\n' >&2 else printf '\nUnable to copy to clipboard; the enhanced prompt is shown above.\n' >&2 fi rm -f "$output_file" command auggie account status } # ----------------------------------------------------------------------------- # Function: plan_build (Cached canonical plan-build launcher) # ----------------------------------------------------------------------------- plan_build() { emulate -L zsh setopt localtraps local base_url="${PLAN_BUILD_BASE_URL:-https://opengist.resetrix.work/weehong/plan-build/raw/HEAD}" local cache_dir="${PLAN_BUILD_CACHE_DIR:-${XDG_CACHE_HOME:-$HOME/.cache}/plan-build}" local curl_command="${PLAN_BUILD_CURL_COMMAND:-curl}" local skill_file="$HOME/.claude/skills/plan-build/SKILL.md" local architect_file="$HOME/.claude/skills/plan-build-architect/SKILL.md" local stage_dir="" release_dir="" current_tmp="" local skill_tmp="" architect_tmp="" skill_backup="" architect_backup="" local activation_lock="$cache_dir/.activate.lock" local activation_lock_fd="" local selected_release="" active_pid="" refresh_rc=1 call_rc=0 interrupted_rc=0 if ! zmodload zsh/system 2>/dev/null; then printf "Error: Unable to load the Zsh system module required by plan-build.\n" >&2 return 1 fi require_cli zsh "Zsh" || return 1 if ! command mkdir -p "$cache_dir" "${skill_file:h}" "${architect_file:h}"; then printf "Error: Unable to create plan-build cache or skill directories.\n" >&2 return 1 fi _plan_build_cleanup() { if [[ -n "${active_pid:-}" ]] && command kill -0 "$active_pid" 2>/dev/null; then command kill "$active_pid" 2>/dev/null || true wait "$active_pid" 2>/dev/null || true fi active_pid="" [[ -n "$stage_dir" && -d "$stage_dir" ]] && command rm -rf -- "$stage_dir" command rm -f -- "$current_tmp" "$skill_tmp" "$architect_tmp" command rm -f -- "$skill_backup" "$architect_backup" if [[ -n "${activation_lock_fd:-}" ]]; then zsystem flock -u "$activation_lock_fd" 2>/dev/null || true activation_lock_fd="" fi } _plan_build_cancel() { interrupted_rc="$1" [[ -n "${active_pid:-}" ]] && command kill "$active_pid" 2>/dev/null || true } trap '_plan_build_cleanup' EXIT trap '_plan_build_cancel 129' HUP trap '_plan_build_cancel 130' INT trap '_plan_build_cancel 143' TERM _plan_build_validate_release() { local candidate="$1" [[ -d "$candidate" && -s "$candidate/plan_build.zsh" && -s "$candidate/SKILL.md" && -s "$candidate/ARCHITECT.md" ]] && command zsh -n "$candidate/plan_build.zsh" } _plan_build_replace_path() { local replacement="$1" destination="$2" if command mv -fT "$replacement" "$destination" 2>/dev/null; then return 0 fi command mv -fh "$replacement" "$destination" } _plan_build_select_current() { local resolved [[ -L "$cache_dir/current" ]] || return 1 resolved="$cache_dir/current" resolved="${resolved:A}" _plan_build_validate_release "$resolved" || return 1 selected_release="$resolved" } _plan_build_acquire_activation_lock() { (( interrupted_rc )) && return 1 command touch "$activation_lock" || return 1 zsystem flock -t 10 -f activation_lock_fd "$activation_lock" } _plan_build_release_activation_lock() { [[ -n "$activation_lock_fd" ]] || return 0 zsystem flock -u "$activation_lock_fd" || return 1 activation_lock_fd="" } _plan_build_download() { local remote="$1" destination="$2" download_rc command "$curl_command" -fsSL "$remote" -o "$destination" & active_pid=$! wait "$active_pid" download_rc=$? active_pid="" (( interrupted_rc )) && return "$interrupted_rc" return "$download_rc" } _plan_build_backup_path() { local source="$1" backup="$2" [[ -e "$source" || -L "$source" ]] || return 0 command rm -f -- "$backup" && command cp -RP "$source" "$backup" } _plan_build_restore_path() { local destination="$1" backup="$2" had_old="$3" if (( had_old )); then _plan_build_replace_path "$backup" "$destination" || return 1 else command rm -f -- "$destination" fi } _plan_build_install_skill_links() { local activation_ok=1 local skill_had_old=0 architect_had_old=0 skill_tmp="$(command mktemp "${skill_file}.link.XXXXXX")" && command rm -f "$skill_tmp" && command ln -s "$cache_dir/current/SKILL.md" "$skill_tmp" || skill_tmp="" architect_tmp="$(command mktemp "${architect_file}.link.XXXXXX")" && command rm -f "$architect_tmp" && command ln -s "$cache_dir/current/ARCHITECT.md" "$architect_tmp" || architect_tmp="" [[ -n "$skill_tmp" && -L "$skill_tmp" && -n "$architect_tmp" && -L "$architect_tmp" ]] || return 1 skill_backup="${skill_file}.backup.$$.$RANDOM" architect_backup="${architect_file}.backup.$$.$RANDOM" [[ -e "$skill_file" || -L "$skill_file" ]] && skill_had_old=1 [[ -e "$architect_file" || -L "$architect_file" ]] && architect_had_old=1 _plan_build_backup_path "$skill_file" "$skill_backup" || activation_ok=0 (( activation_ok )) && _plan_build_backup_path "$architect_file" "$architect_backup" || activation_ok=0 (( activation_ok && ! interrupted_rc )) && _plan_build_replace_path "$skill_tmp" "$skill_file" || activation_ok=0 (( activation_ok )) && skill_tmp="" (( activation_ok && ! interrupted_rc )) && _plan_build_replace_path "$architect_tmp" "$architect_file" || activation_ok=0 (( activation_ok )) && architect_tmp="" if (( activation_ok )); then command rm -f -- "$skill_backup" "$architect_backup" skill_backup="" architect_backup="" return 0 fi if ! _plan_build_restore_path "$skill_file" "$skill_backup" "$skill_had_old"; then printf "Error: Unable to restore plan-build skill; backup preserved at %s.\n" "$skill_backup" >&2 skill_backup="" fi if ! _plan_build_restore_path "$architect_file" "$architect_backup" "$architect_had_old"; then printf "Error: Unable to restore architect skill; backup preserved at %s.\n" "$architect_backup" >&2 architect_backup="" fi return 1 } if (( $+commands[$curl_command] )) || [[ "$curl_command" == */* && -x "$curl_command" ]]; then stage_dir="$(command mktemp -d "$cache_dir/.stage.XXXXXX")" || stage_dir="" if [[ -n "$stage_dir" ]]; then _plan_build_download "$base_url/plan_build.zsh" "$stage_dir/plan_build.zsh" && _plan_build_download "$base_url/SKILL.md" "$stage_dir/SKILL.md" && _plan_build_download "$base_url/ARCHITECT.md" "$stage_dir/ARCHITECT.md" refresh_rc=$? if (( interrupted_rc )); then _plan_build_cleanup trap - EXIT HUP INT TERM return "$interrupted_rc" fi else refresh_rc=1 fi if (( refresh_rc == 0 )) && _plan_build_validate_release "$stage_dir"; then release_dir="$(command mktemp -d "$cache_dir/release.XXXXXX")" || release_dir="" if [[ -n "$release_dir" ]] && command rmdir "$release_dir" && command mv "$stage_dir" "$release_dir" && _plan_build_acquire_activation_lock; then stage_dir="" release_dir="${release_dir:A}" current_tmp="$(command mktemp "$cache_dir/.current.XXXXXX")" && command rm -f "$current_tmp" && command ln -s "$release_dir" "$current_tmp" || current_tmp="" if [[ -n "$current_tmp" && -L "$current_tmp" ]] && _plan_build_install_skill_links && _plan_build_replace_path "$current_tmp" "$cache_dir/current"; then current_tmp="" selected_release="$release_dir" fi _plan_build_release_activation_lock || true fi fi fi if (( interrupted_rc )); then _plan_build_cleanup trap - EXIT HUP INT TERM return "$interrupted_rc" fi if [[ -z "$selected_release" ]]; then if ! _plan_build_select_current; then if (( ! $+commands[$curl_command] )) && [[ ! ( "$curl_command" == */* && -x "$curl_command" ) ]]; then printf "Error: curl is required because no valid cached plan-build bundle is available.\n" >&2 else printf "Error: Unable to refresh plan-build and no valid cached bundle is available.\n" >&2 fi return 1 fi if (( ! $+commands[$curl_command] )) && [[ ! ( "$curl_command" == */* && -x "$curl_command" ) ]]; then printf "Warning: curl is unavailable; using the validated cached plan-build bundle.\n" >&2 else printf "Warning: Unable to refresh plan-build; using the validated cached bundle.\n" >&2 fi fi if [[ "$(command readlink "$skill_file" 2>/dev/null)" != "$cache_dir/current/SKILL.md" || "$(command readlink "$architect_file" 2>/dev/null)" != "$cache_dir/current/ARCHITECT.md" ]]; then if ! _plan_build_acquire_activation_lock || ! _plan_build_install_skill_links; then _plan_build_release_activation_lock 2>/dev/null || true if (( interrupted_rc )); then _plan_build_cleanup trap - EXIT HUP INT TERM return "$interrupted_rc" fi printf "Error: Unable to install the plan-build skill links.\n" >&2 return 1 fi _plan_build_release_activation_lock || true fi if (( interrupted_rc )); then _plan_build_cleanup trap - EXIT HUP INT TERM return "$interrupted_rc" fi PLAN_BUILD_SKILL_PATH="$selected_release/SKILL.md" \ PLAN_BUILD_ARCHITECT_SKILL_PATH="$selected_release/ARCHITECT.md" \ command zsh "$selected_release/plan_build.zsh" "$@" & active_pid=$! wait "$active_pid" call_rc=$? active_pid="" (( interrupted_rc )) && call_rc="$interrupted_rc" _plan_build_cleanup trap - EXIT HUP INT TERM return "$call_rc" } # ----------------------------------------------------------------------------- # Internal: _aicommit_usage (Print aicommit usage) # ----------------------------------------------------------------------------- _aicommit_usage() { printf "Usage: aicommit [--modules] [--yes] [claude | copilot | opencode]\n" printf " --modules Force per-module commits (automatic above 30 staged paths)\n" printf " --yes Skip the plan confirmation for per-module commits\n" } # ----------------------------------------------------------------------------- # Internal: _aicommit_staged_diff (Staged diff without lock/minified files) # ----------------------------------------------------------------------------- _aicommit_staged_diff() { # Exclude lock files/minified files and cap at ~100k characters to prevent CLI crashes git diff --cached -- . ":(exclude)*.lock" ":(exclude)*-lock.json" ":(exclude)*.min.js" | head -c 100000 } # ----------------------------------------------------------------------------- # Internal: _aicommit_opencode_model (Read OpenCode's most recent model selection) # ----------------------------------------------------------------------------- _aicommit_opencode_model() { python3 - <<'PY' import json import os from pathlib import Path state = Path(os.environ.get("XDG_STATE_HOME") or Path.home() / ".local/state") try: data = json.loads((state / "opencode/model.json").read_text()) recent = data.get("recent", []) if recent: selected = recent[0] model = selected["providerID"] + "/" + selected["modelID"] variant = data.get("variant", {}).get(model) if variant and variant != "default": model += "#" + variant print(model) except FileNotFoundError: pass # Let OpenCode use its configured default on a fresh installation. PY } # ----------------------------------------------------------------------------- # Internal: _aicommit_engine (Ask an AI CLI to commit or to plan a split) # ----------------------------------------------------------------------------- _aicommit_engine() { local engine="$1" local mode="$2" local input="$3" local name local selected_model local -a model_args model_args=() if [ "$engine" = "opencode" ]; then selected_model="$(_aicommit_opencode_model)" || return 1 if [ -n "$selected_model" ]; then model_args=(--model "$selected_model") fi fi case "$engine" in claude) name="Claude" ;; copilot) name="Copilot" ;; opencode) name="OpenCode" ;; esac # Status goes to stderr so plan output on stdout stays clean JSON. if [ "$mode" = "plan" ]; then printf "🧩 %s is planning per-module commits...\n" "$name" >&2 else printf "🤖 %s is analyzing staged changes...\n" "$name" >&2 fi case "$engine:$mode" in claude:commit) printf "%s\n" "$input" \ | command claude --allowedTools 'Bash(git commit *)' --permission-mode dontAsk -p \ "Read the provided instructions and diff, then generate and execute the commit." ;; claude:plan) printf "%s\n" "$input" \ | command claude --permission-mode dontAsk -p \ "Read the provided instructions and staged changes, then reply with only the JSON plan." ;; copilot:commit) copilot -p "$input" -s --no-ask-user --allow-tool='shell(git:*)' ;; copilot:plan) copilot -p "$input" -s --no-ask-user ;; opencode:commit) printf "%s\n" "$input" | opencode run --agent ai-commit "${model_args[@]}" ;; opencode:plan) printf "%s\n" "$input" | opencode run --agent ai-commit-plan "${model_args[@]}" ;; esac } # ----------------------------------------------------------------------------- # Internal: _aicommit_restage (Stage HEAD plus plan groups first..last from a tree) # ----------------------------------------------------------------------------- _aicommit_restage() { local tree="$1" local run_dir="$2" local first="$3" local last="$4" local number if git rev-parse -q --verify HEAD >/dev/null; then git read-tree HEAD || return 1 else git read-tree --empty || return 1 fi [ "$first" -le "$last" ] || return 0 number="$first" while [ "$number" -le "$last" ]; do cat "$run_dir/$(printf "%03d" "$number").files" number=$((number + 1)) done | git --literal-pathspecs restore --staged --source="$tree" --pathspec-from-file=- --pathspec-file-nul } # ----------------------------------------------------------------------------- # Internal: _aicommit_plan (Validate the AI split plan into per-group files) # ----------------------------------------------------------------------------- _aicommit_plan() { local run_dir="$1" python3 -c ' import json, os, re, sys sys.stdout.reconfigure(errors="backslashreplace") sys.stderr.reconfigure(errors="backslashreplace") run_dir = sys.argv[1] types = {"feat", "fix", "docs", "style", "refactor", "perf", "test", "build", "ci", "chore"} with open(os.path.join(run_dir, "staged"), "rb") as fh: staged = [p for p in fh.read().decode("utf-8", "surrogateescape").split("\0") if p] with open(os.path.join(run_dir, "plan.out"), encoding="utf-8", errors="replace") as fh: raw = re.sub(r"\x1b\[[0-?]*[ -/]*[@-~]", "", fh.read()) # Take the last {"commits": [...]} object in the reply, ignoring fences and prose. plan, decoder, pos = None, json.JSONDecoder(), raw.find("{") while pos != -1: try: obj, end = decoder.raw_decode(raw, pos) except ValueError: pos = raw.find("{", pos + 1) continue if isinstance(obj, dict) and isinstance(obj.get("commits"), list): plan = obj pos = raw.find("{", end) else: pos = raw.find("{", pos + 1) if plan is None: sys.exit("❌ Error: The AI reply did not contain a JSON commit plan.") # Every staged path must land in exactly one group; anything unassigned goes last. staged_set, seen, groups, warnings = set(staged), set(), [], [] for commit in plan["commits"]: if not isinstance(commit, dict) or not isinstance(commit.get("files"), list): continue files = [] for path in commit["files"]: if not isinstance(path, str): continue if path not in staged_set: warnings.append("ignoring path that is not staged: " + path) elif path in seen: warnings.append("ignoring duplicate path: " + path) else: seen.add(path) files.append(path) if not files: continue kind = str(commit.get("type", "")).strip().lower() scope = re.sub(r"[^a-z0-9._/-]", "", str(commit.get("scope", "")).lower()) label = "" if kind in types: label = kind + ("(" + scope + ")" if scope else "") groups.append((label, files)) missing = [p for p in staged if p not in seen] if missing: warnings.append("%d staged path(s) were not planned; adding them as a final commit" % len(missing)) groups.append(("", missing)) if len(staged) > 30 and len(groups) < 2: sys.exit("❌ Error: Changes affecting more than 30 staged paths require multiple module commits.") for number, (label, files) in enumerate(groups, 1): base = os.path.join(run_dir, "%03d" % number) with open(base + ".files", "wb") as fh: fh.write("".join(p + "\0" for p in files).encode("utf-8", "surrogateescape")) with open(base + ".label", "w") as fh: fh.write(label) with open(os.path.join(run_dir, "count"), "w") as fh: fh.write(str(len(groups))) for warning in warnings: print("⚠️ " + warning, file=sys.stderr) print("📋 Planned commits:") for number, (label, files) in enumerate(groups, 1): print(" %d. %s (%d file%s)" % (number, label or "unplanned", len(files), "" if len(files) == 1 else "s")) for path in files: print(" " + path) ' "$run_dir" } # ----------------------------------------------------------------------------- # Internal: _aicommit_modules_run (Plan, confirm, and commit each module group) # ----------------------------------------------------------------------------- _aicommit_modules_run() { local engine="$1" local assume_yes="$2" local run_dir="$3" local prompt_file="$HOME/.config/ai-commit-prompt.txt" local split_prompt_file="$HOME/.config/ai-commit-split-prompt.txt" local tree local total local current local group local label local hint local diff local expected local head_before local head_after local unexpected local reply if ! tree="$(git write-tree 2>/dev/null)"; then printf "❌ Error: Unable to snapshot the index (unresolved conflicts?).\n" >&2 return 1 fi git diff --cached --name-only --no-renames -z > "$run_dir/staged" || return 1 if [ ! -s "$run_dir/staged" ]; then printf "❌ Error: Nothing staged to commit.\n" >&2 return 1 fi _aicommit_engine "$engine" plan "$(cat "$split_prompt_file") Here are the staged files: $(git -c core.quotePath=false diff --cached --name-status -M) Here is the diffstat: $(git -c core.quotePath=false diff --cached --numstat -M) Here is the git diff: $(_aicommit_staged_diff)" > "$run_dir/plan.out" if ! _aicommit_plan "$run_dir"; then printf "Last lines of the AI reply:\n" >&2 tail -n 20 "$run_dir/plan.out" >&2 return 1 fi total="$(cat "$run_dir/count")" if [ "$assume_yes" -ne 1 ]; then if ! (: </dev/tty) 2>/dev/null; then printf "❌ Error: No interactive terminal to confirm the plan; rerun with --yes.\n" >&2 return 1 fi printf "Create these %s commit(s)? (Y/n) " "$total" >/dev/tty IFS= read -r reply </dev/tty || reply="n" case "$reply" in ""|y|Y|yes|YES|Yes) ;; *) printf "Aborted. Staged changes are untouched.\n" return 1 ;; esac fi printf "💾 Staged snapshot: %s (recover with: git read-tree %s)\n" "$tree" "$tree" current=1 trap '_aicommit_restage "$tree" "$run_dir" "$current" "$total"; printf "\n⚠️ Interrupted. Uncommitted groups are staged again.\n" >&2; rm -rf "$run_dir"; exit 130' INT TERM while [ "$current" -le "$total" ]; do group="$run_dir/$(printf "%03d" "$current").files" label="$(cat "$run_dir/$(printf "%03d" "$current").label")" if ! _aicommit_restage "$tree" "$run_dir" "$current" "$current"; then printf "❌ Error: Unable to stage commit %s/%s.\n" "$current" "$total" >&2 _aicommit_restage "$tree" "$run_dir" "$current" "$total" return 1 fi expected="$(git write-tree)" head_before="$(git rev-parse -q --verify HEAD)" diff="$(_aicommit_staged_diff)" [ -n "$diff" ] || diff="$(git diff --cached --stat)" hint="This is commit $current of $total, split by module." if [ -n "$label" ]; then hint="$hint The planned type and scope is $label; use it unless the diff clearly calls for another." fi printf "\n[%s/%s] %s\n" "$current" "$total" "${label:-unplanned}" _aicommit_engine "$engine" commit "$(cat "$prompt_file") $hint Here is the git diff: $diff" head_after="$(git rev-parse -q --verify HEAD)" if [ -z "$head_after" ] || [ "$head_after" = "$head_before" ]; then printf "❌ Error: No commit was created for group %s/%s.\n" "$current" "$total" >&2 _aicommit_restage "$tree" "$run_dir" "$current" "$total" printf "Uncommitted groups are staged again.\n" >&2 return 1 fi # Hooks may reformat this group's files, but nothing outside it may sneak in. unexpected="$(LC_ALL=C comm -13 \ <(tr '\0' '\n' < "$group" | LC_ALL=C sort) \ <(git diff --name-only --no-renames -z "$expected" HEAD | tr '\0' '\n' | LC_ALL=C sort))" if [ -n "$unexpected" ]; then printf "❌ Error: Commit %s/%s included files outside its group:\n%s\n" "$current" "$total" "$unexpected" >&2 _aicommit_restage "$tree" "$run_dir" "$((current + 1))" "$total" printf "Inspect it with 'git show'; undo it with 'git reset --soft HEAD~1'.\n" >&2 return 1 fi current=$((current + 1)) done trap - INT TERM printf "\n✅ Created %s commit(s):\n" "$total" git --no-pager log --oneline -n "$total" } # ----------------------------------------------------------------------------- # Internal: _aicommit_modules (Split staged changes into per-module commits) # ----------------------------------------------------------------------------- _aicommit_modules() { local engine="$1" local assume_yes="$2" local top local run_dir local rc if [ ! -f "$HOME/.config/ai-commit-split-prompt.txt" ]; then printf "❌ Error: Configuration file not found at ~/.config/ai-commit-split-prompt.txt\n" >&2 return 1 fi require_cli python3 "Python 3" || return 1 if ! top="$(git rev-parse --show-toplevel 2>/dev/null)"; then printf "❌ Error: Not inside a git work tree.\n" >&2 return 1 fi run_dir="$(mktemp -d "${TMPDIR:-/tmp}/aicommit.XXXXXX")" || return 1 # Subshell: run from the repo root and keep the traps away from the caller. ( trap 'rm -rf "$run_dir"; exit 130' INT TERM cd "$top" || exit 1 _aicommit_modules_run "$engine" "$assume_yes" "$run_dir" ) rc=$? rm -rf "$run_dir" return "$rc" } # ----------------------------------------------------------------------------- # Function: aicommit (Automated Conventional Commit Engine Wrapper) # ----------------------------------------------------------------------------- aicommit() { local engine="claude" local modules=0 local assume_yes=0 local arg local diff local staged_count local prompt_file="$HOME/.config/ai-commit-prompt.txt" for arg in "$@"; do case "$arg" in claude|copilot|opencode) engine="$arg" ;; --modules) modules=1 ;; -y|--yes) assume_yes=1 ;; -h|--help) _aicommit_usage return 0 ;; *) printf "❌ Error: Invalid argument '%s'\n" "$arg" >&2 _aicommit_usage >&2 return 1 ;; esac done if [ ! -f "$prompt_file" ]; then printf "❌ Error: Configuration file not found at ~/.config/ai-commit-prompt.txt\n" >&2 printf "Please create it and paste your system prompt inside.\n" >&2 return 1 fi # Count the full index, including excluded previews and both sides of renames. # NUL delimiters keep whitespace/newlines in filenames from skewing the count. staged_count="$(set -o pipefail; git diff --cached --name-only --no-renames -z | tr -cd '\0' | wc -c)" || return 1 if [ "$staged_count" -gt 30 ] && [ "$modules" -ne 1 ]; then printf "🧩 %s staged paths exceed 30; splitting commits by module.\n" "$staged_count" >&2 fi if [ "$modules" -eq 1 ] || [ "$staged_count" -gt 30 ]; then _aicommit_modules "$engine" "$assume_yes" return fi diff="$(_aicommit_staged_diff)" if [ -z "$diff" ]; then printf "❌ Error: Nothing staged to commit.\n" >&2 return 1 fi _aicommit_engine "$engine" commit "$(cat "$prompt_file") Staged path count: $staged_count Here is the git diff: $diff" } # ----------------------------------------------------------------------------- # Function: create (Efficiently create file + parent dirs) # ----------------------------------------------------------------------------- create() { if ! command -v install >/dev/null 2>&1; then printf "Error: 'install' coreutil is not available.\n" >&2 return 1 fi if [ $# -eq 0 ]; then printf "Usage: create <file> [file ...]\n" >&2 return 2 fi for p in "$@"; do # Expand tilde manually if shell doesn't [[ "$p" == "~/"* ]] && p="${HOME}/${p#\~/}" if [ -z "$p" ] || [ "$p" = "/" ]; then printf "create: refusing to operate on '%s'\n" "$p" >&2 continue fi if [ -e "$p" ]; then printf "create: '%s' already exists. Skipping.\n" "$p" >&2 continue fi if install -D /dev/null "$p"; then echo "Created '$p'" fi done } # ----------------------------------------------------------------------------- # Function: ckp (Kill or check the process using a port) # ----------------------------------------------------------------------------- ckp() { local action="kill" local signal="TERM" local port local pids local pid case "${1:-}" in -c|--check) action="check" shift ;; -k|--kill) action="kill" shift ;; -f|--force) action="kill" signal="KILL" shift ;; -h|--help|"") printf "Usage: ckp [--check|-c|--kill|-k|--force|-f] <port_number>\n" >&2 return 1 ;; esac port="${1:-}" if [[ ! "$port" =~ ^[0-9]+$ ]] || [ "$port" -lt 1 ] || [ "$port" -gt 65535 ]; then printf "Error: port must be a number between 1 and 65535.\n" >&2 printf "Usage: ckp [--check|-c|--kill|-k|--force|-f] <port_number>\n" >&2 return 1 fi if [[ "$OSTYPE" == "darwin"* ]]; then pids="$( { lsof -nP -tiTCP:"$port" -sTCP:LISTEN 2>/dev/null lsof -nP -tiUDP:"$port" 2>/dev/null } | sort -u )" if [ -z "$pids" ]; then printf "Port %s is FREE\n" "$port" return 0 fi printf "Port %s is in use:\n" "$port" lsof -nP -iTCP:"$port" -sTCP:LISTEN -iUDP:"$port" 2>/dev/null else local port_hex local inodes port_hex="$(printf "%04X" "$port")" inodes="$( awk -v port="$port_hex" ' NR > 1 { split($2, local_addr, ":") if (toupper(local_addr[length(local_addr)]) == port && ($4 == "0A" || FILENAME ~ /udp/)) { print $10 } } ' /proc/net/tcp /proc/net/tcp6 /proc/net/udp /proc/net/udp6 2>/dev/null | sort -u )" if [ -z "$inodes" ]; then printf "Port %s is FREE\n" "$port" return 0 fi pids="" pids="$( find /proc/[0-9]*/fd -maxdepth 1 -type l -printf "%p %l\n" 2>/dev/null | awk -v inodes="$inodes" ' BEGIN { split(inodes, inode_list, " ") for (i in inode_list) { wanted["socket:[" inode_list[i] "]"] = 1 } } $2 in wanted { split($1, path, "/") print path[3] } ' | sort -u )" if [ -z "$pids" ]; then printf "Port %s is in use, but no readable process owner was found.\n" "$port" >&2 return 1 fi printf "Port %s is in use:\n" "$port" for pid in $pids; do if [ -r "/proc/$pid/cmdline" ]; then printf " PID %-8s %s\n" "$pid" "$(tr '\0' ' ' < "/proc/$pid/cmdline")" else printf " PID %s\n" "$pid" fi done fi if [ "$action" = "kill" ]; then for pid in $pids; do if kill "-$signal" "$pid" 2>/dev/null; then printf "Killed PID %s with SIG%s.\n" "$pid" "$signal" else printf "Error: failed to kill PID %s. Try with sudo or --force.\n" "$pid" >&2 return 1 fi done fi } # ----------------------------------------------------------------------------- # Function: append_path (Adds to session PATH if exists) # ----------------------------------------------------------------------------- append_path() { local dir="$1" if [ -d "$dir" ]; then case ":$PATH:" in *":$dir:"*) ;; *) PATH="$PATH:$dir" ;; esac elif [ "${SUPPRESS_WARNINGS:-0}" -ne 1 ]; then echo "Warning: $dir does not exist." >&2 fi } # ----------------------------------------------------------------------------- # Function: add_path_to_config (Fixed awk/sed injection) # ----------------------------------------------------------------------------- add_path_to_config() { local new_path="$1" local config_file="$HOME/.bashrc" [[ "$SHELL" == *"zsh"* ]] && config_file="$HOME/.zshrc" if [ -z "$new_path" ]; then printf "Usage: add_path_to_config <directory>\n" >&2 return 1 fi new_path="${new_path/#\~/$HOME}" if [ ! -d "$new_path" ]; then printf "Warning: '%s' does not exist. Add anyway? (y/N) " "$new_path" read -r REPLY [[ ! $REPLY =~ ^[Yy]$ ]] && return 1 fi # Check if already exists in file if grep -Fq "append_path \"$new_path\"" "$config_file"; then printf "Notice: Already in %s.\n" "$config_file" return 0 fi # Robust Injection: Use sed to insert before export PATH or just append if grep -q "^export PATH" "$config_file"; then sed -i.bak "/^export PATH/i append_path \"$new_path\"" "$config_file" else echo "append_path \"$new_path\"" >> "$config_file" echo "export PATH" >> "$config_file" fi printf "✅ Success: Added to %s\n" "$config_file" append_path "$new_path" && export PATH }

# ============================================================================= # ENVIRONMENT & PATH CONFIGURATION # ============================================================================= # 1. Source functions first if [[ -f "$HOME/.func" && -r "$HOME/.func" ]]; then source "$HOME/.func" fi # 2. Define Root Variables export DOTNET_ROOT="$HOME/.dotnet" export PNPM_HOME="$HOME/.local/share/pnpm" # 3. PATH INITIALIZATION # ----------------------------------------------------------------------------- append_path "$HOME/.local/bin" append_path "$DOTNET_ROOT" append_path "$DOTNET_ROOT/tools" append_path "$HOME/.opencode/bin" append_path "$HOME/Flutter/bin" append_path "$PNPM_HOME/bin" # macOS specific paths if [[ "$OSTYPE" == "darwin"* ]]; then append_path "/Applications/Espanso.app/Contents/MacOS" fi # Finalize PATH export PATH

# ============================================================================= # FILE: ~/.sourcerc # Description: Initializes third-party package managers and external tools. # ============================================================================= # ============================================================================= # GENERAL ENV # ============================================================================= if [[ -f "$HOME/.local/bin/env" && -r "$HOME/.local/bin/env" ]]; then source "$HOME/.local/bin/env" fi # ============================================================================= # SDKMAN (Java/Kotlin/Scala Version Manager) # ============================================================================= export SDKMAN_DIR="$HOME/.sdkman" if [[ -f "$SDKMAN_DIR/bin/sdkman-init.sh" && -r "$SDKMAN_DIR/bin/sdkman-init.sh" ]]; then source "$SDKMAN_DIR/bin/sdkman-init.sh" fi # ============================================================================= # NVM (Node Version Manager) # ============================================================================= export NVM_DIR="$HOME/.nvm" if [[ -f "$NVM_DIR/nvm.sh" && -r "$NVM_DIR/nvm.sh" ]]; then source "$NVM_DIR/nvm.sh" fi if [[ -f "$NVM_DIR/bash_completion" && -r "$NVM_DIR/bash_completion" ]]; then source "$NVM_DIR/bash_completion" fi # ============================================================================= # GOOGLE CLOUD SDK # ============================================================================= if [[ -f "/opt/google-cloud-sdk/path.zsh.inc" ]]; then source "/opt/google-cloud-sdk/path.zsh.inc" fi if [[ -f "/opt/google-cloud-sdk/completion.zsh.inc" ]]; then source "/opt/google-cloud-sdk/completion.zsh.inc" fi

"$schema" = 'https://starship.rs/config-schema.json' format = """ [](red)\ $os\ $username\ [](bg:peach fg:red)\ $directory\ [](bg:yellow fg:peach)\ $git_branch\ $git_status\ [](fg:yellow bg:green)\ $c\ $rust\ $golang\ $nodejs\ $bun\ $php\ $java\ $kotlin\ $haskell\ $python\ [](fg:green bg:sapphire)\ $conda\ [](fg:sapphire bg:lavender)\ $time\ [ ](fg:lavender)\ $cmd_duration\ $line_break\ $character""" palette = 'catppuccin_mocha' [os] disabled = false style = "bg:red fg:crust" [os.symbols] Windows = "" Ubuntu = "󰕈" SUSE = "" Raspbian = "󰐿" Mint = "󰣭" Macos = "󰀵" Manjaro = "" Linux = "󰌽" Gentoo = "󰣨" Fedora = "󰣛" Alpine = "" Amazon = "" Android = "" AOSC = "" Arch = "󰣇" Artix = "󰣇" CentOS = "" Debian = "󰣚" Redhat = "󱄛" RedHatEnterprise = "󱄛" [username] show_always = true style_user = "bg:red fg:crust" style_root = "bg:red fg:crust" format = '[ $user]($style)' [directory] style = "bg:peach fg:crust" format = "[ $path ]($style)" truncation_length = 3 truncation_symbol = "…/" [directory.substitutions] "Documents" = "󰈙 " "Downloads" = " " "Music" = "󰝚 " "Pictures" = " " "Developer" = "󰲋 " [git_branch] symbol = "" style = "bg:yellow" format = '[[ $symbol $branch ](fg:crust bg:yellow)]($style)' [git_status] style = "bg:yellow" format = '[[($all_status$ahead_behind )](fg:crust bg:yellow)]($style)' [nodejs] symbol = "" style = "bg:green" format = '[[ $symbol( $version) ](fg:crust bg:green)]($style)' [bun] symbol = "" style = "bg:green" format = '[[ $symbol( $version) ](fg:crust bg:green)]($style)' [c] symbol = " " style = "bg:green" format = '[[ $symbol( $version) ](fg:crust bg:green)]($style)' [rust] symbol = "" style = "bg:green" format = '[[ $symbol( $version) ](fg:crust bg:green)]($style)' [golang] symbol = "" style = "bg:green" format = '[[ $symbol( $version) ](fg:crust bg:green)]($style)' [php] symbol = "" style = "bg:green" format = '[[ $symbol( $version) ](fg:crust bg:green)]($style)' [java] symbol = " " style = "bg:green" format = '[[ $symbol( $version) ](fg:crust bg:green)]($style)' [kotlin] symbol = "" style = "bg:green" format = '[[ $symbol( $version) ](fg:crust bg:green)]($style)' [haskell] symbol = "" style = "bg:green" format = '[[ $symbol( $version) ](fg:crust bg:green)]($style)' [python] symbol = "" style = "bg:green" format = '[[ $symbol( $version)(\(#$virtualenv\)) ](fg:crust bg:green)]($style)' [docker_context] symbol = "" style = "bg:sapphire" format = '[[ $symbol( $context) ](fg:crust bg:sapphire)]($style)' [conda] symbol = "  " style = "fg:crust bg:sapphire" format = '[$symbol$environment ]($style)' ignore_base = false [time] disabled = false time_format = "%R" style = "bg:lavender" format = '[[  $time ](fg:crust bg:lavender)]($style)' [line_break] disabled = false [character] disabled = false success_symbol = '[❯](bold fg:green)' error_symbol = '[❯](bold fg:red)' vimcmd_symbol = '[❮](bold fg:green)' vimcmd_replace_one_symbol = '[❮](bold fg:lavender)' vimcmd_replace_symbol = '[❮](bold fg:lavender)' vimcmd_visual_symbol = '[❮](bold fg:yellow)' [cmd_duration] show_milliseconds = true format = " in $duration " style = "bg:lavender" disabled = false show_notifications = true min_time_to_notify = 45000 [palettes.catppuccin_mocha] rosewater = "#f5e0dc" flamingo = "#f2cdcd" pink = "#f5c2e7" mauve = "#cba6f7" red = "#f38ba8" maroon = "#eba0ac" peach = "#fab387" yellow = "#f9e2af" green = "#a6e3a1" teal = "#94e2d5" sky = "#89dceb" sapphire = "#74c7ec" blue = "#89b4fa" lavender = "#b4befe" text = "#cdd6f4" subtext1 = "#bac2de" subtext0 = "#a6adc8" overlay2 = "#9399b2" overlay1 = "#7f849c" overlay0 = "#6c7086" surface2 = "#585b70" surface1 = "#45475a" surface0 = "#313244" base = "#1e1e2e" mantle = "#181825" crust = "#11111b" [palettes.catppuccin_frappe] rosewater = "#f2d5cf" flamingo = "#eebebe" pink = "#f4b8e4" mauve = "#ca9ee6" red = "#e78284" maroon = "#ea999c" peach = "#ef9f76" yellow = "#e5c890" green = "#a6d189" teal = "#81c8be" sky = "#99d1db" sapphire = "#85c1dc" blue = "#8caaee" lavender = "#babbf1" text = "#c6d0f5" subtext1 = "#b5bfe2" subtext0 = "#a5adce" overlay2 = "#949cbb" overlay1 = "#838ba7" overlay0 = "#737994" surface2 = "#626880" surface1 = "#51576d" surface0 = "#414559" base = "#303446" mantle = "#292c3c" crust = "#232634" [palettes.catppuccin_latte] rosewater = "#dc8a78" flamingo = "#dd7878" pink = "#ea76cb" mauve = "#8839ef" red = "#d20f39" maroon = "#e64553" peach = "#fe640b" yellow = "#df8e1d" green = "#40a02b" teal = "#179299" sky = "#04a5e5" sapphire = "#209fb5" blue = "#1e66f5" lavender = "#7287fd" text = "#4c4f69" subtext1 = "#5c5f77" subtext0 = "#6c6f85" overlay2 = "#7c7f93" overlay1 = "#8c8fa1" overlay0 = "#9ca0b0" surface2 = "#acb0be" surface1 = "#bcc0cc" surface0 = "#ccd0da" base = "#eff1f5" mantle = "#e6e9ef" crust = "#dce0e8" [palettes.catppuccin_macchiato] rosewater = "#f4dbd6" flamingo = "#f0c6c6" pink = "#f5bde6" mauve = "#c6a0f6" red = "#ed8796" maroon = "#ee99a0" peach = "#f5a97f" yellow = "#eed49f" green = "#a6da95" teal = "#8bd5ca" sky = "#91d7e3" sapphire = "#7dc4e4" blue = "#8aadf4" lavender = "#b7bdf8" text = "#cad3f5" subtext1 = "#b8c0e0" subtext0 = "#a5adcb" overlay2 = "#939ab7" overlay1 = "#8087a2" overlay0 = "#6e738d" surface2 = "#5b6078" surface1 = "#494d64" surface0 = "#363a4f" base = "#24273a" mantle = "#1e2030" crust = "#181926"

#!/usr/bin/env python3 """Exercise aicommit routing and plan validation without invoking an AI CLI.""" import json import os from pathlib import Path import subprocess import tempfile import unittest SOURCE = Path(__file__).resolve().parent / "func" ROUTE = ''' source "$1" shift _aicommit_engine() { printf 'ENGINE %s %s\n' "$1" "$2"; } _aicommit_modules() { printf 'MODULES %s %s\n' "$1" "$2"; return 7; } aicommit "$@" ''' class AicommitTests(unittest.TestCase): def setUp(self): self.temp = tempfile.TemporaryDirectory(prefix="aicommit-test-") self.addCleanup(self.temp.cleanup) self.root = Path(self.temp.name) self.repo = self.root / "repo" self.repo.mkdir() self.home = self.root / "home" config = self.home / ".config" config.mkdir(parents=True) for name in ("ai-commit-prompt.txt", "ai-commit-split-prompt.txt"): (config / name).write_text((SOURCE.parent / name).read_text()) self.env = {**os.environ, "HOME": str(self.home)} self.git("init", "-q") def git(self, *args): return subprocess.check_output( ["git", *args], cwd=self.repo, env=self.env ) def shell(self, script, *args, cwd=None): return subprocess.run( ["zsh", "-f", "-c", script, "test", str(SOURCE), *args], cwd=cwd or self.repo, env=self.env, text=True, capture_output=True, ) def stage(self, count, suffix=".txt"): # A newline is part of one path, not an extra staged file. files = [f"file {i}{suffix}" for i in range(count)] if files: files[0] = "line\nbreak" + suffix for name in files: (self.repo / name).write_text("staged content\n") self.git("add", "--all") return files def test_thirty_paths_use_single_commit(self): self.stage(30) result = self.shell(ROUTE, "opencode") self.assertEqual(result.returncode, 0, result.stderr) self.assertEqual(result.stdout.strip(), "ENGINE opencode commit") def test_thirty_one_paths_route_all_engines_and_forward_yes(self): self.stage(31) for engine in ("claude", "copilot", "opencode"): with self.subTest(engine=engine): result = self.shell(ROUTE, "--yes", engine) self.assertEqual(result.returncode, 7, result.stderr) self.assertEqual(result.stdout.strip(), f"MODULES {engine} 1") self.assertIn("exceed 30", result.stderr) def test_explicit_modules_below_threshold(self): self.stage(1) result = self.shell(ROUTE, "--modules", "copilot") self.assertEqual(result.returncode, 7, result.stderr) self.assertEqual(result.stdout.strip(), "MODULES copilot 0") def test_filtered_files_still_trigger_modules_from_subdirectory(self): self.stage(31, ".lock") nested = self.repo / "nested" nested.mkdir() result = self.shell(ROUTE, cwd=nested) self.assertEqual(result.returncode, 7, result.stderr) self.assertEqual(result.stdout.strip(), "MODULES claude 0") def test_empty_index_and_git_errors_never_call_engine(self): for cwd in (self.repo, self.root): with self.subTest(cwd=cwd): result = self.shell(ROUTE, cwd=cwd) self.assertNotEqual(result.returncode, 0) self.assertNotIn("ENGINE", result.stdout) self.assertNotIn("MODULES", result.stdout) def plan(self, count, groups): run = self.root / "plan" run.mkdir() files = [f"file-{i}" for i in range(count)] (run / "staged").write_bytes(b"".join(p.encode() + b"\0" for p in files)) (run / "plan.out").write_text(json.dumps({"commits": [ {"type": "feat", "scope": "test", "files": group(files)} for group in groups ]})) return self.shell('source "$1"; _aicommit_plan "$2"', str(run)), run def test_single_group_plan_allowed_at_thirty(self): result, run = self.plan(30, [lambda files: files]) self.assertEqual(result.returncode, 0, result.stderr) self.assertEqual((run / "count").read_text(), "1") def test_single_group_plan_rejected_above_thirty_after_deduplication(self): result, run = self.plan(31, [lambda files: files, lambda files: files]) self.assertNotEqual(result.returncode, 0) self.assertIn("require multiple module commits", result.stderr) self.assertFalse((run / "count").exists()) def test_empty_plan_cannot_fall_back_to_one_large_commit(self): result, _ = self.plan(31, []) self.assertNotEqual(result.returncode, 0) self.assertIn("require multiple module commits", result.stderr) def test_multiple_groups_allowed_even_with_large_module(self): result, run = self.plan(32, [lambda files: files[:-1], lambda files: files[-1:]]) self.assertEqual(result.returncode, 0, result.stderr) self.assertEqual((run / "count").read_text(), "2") def test_rejected_auto_plan_preserves_partial_index_and_worktree(self): files = self.stage(31) changed = self.repo / files[0] changed.write_text("unstaged content\n") before = self.git("write-tree") plan = self.root / "single-plan.json" plan.write_text(json.dumps({"commits": [{"files": files}]})) result = self.shell(''' source "$1" plan_file="$2" _aicommit_engine() { if [ "$2" = plan ]; then cat "$plan_file"; else printf 'UNEXPECTED COMMIT\n'; fi } aicommit --yes opencode ''', str(plan)) self.assertNotEqual(result.returncode, 0) self.assertIn("require multiple module commits", result.stderr) self.assertNotIn("UNEXPECTED COMMIT", result.stdout) self.assertEqual(self.git("write-tree"), before) self.assertEqual(changed.read_text(), "unstaged content\n") if __name__ == "__main__": unittest.main()