state-directory-manager
Manage persistent state directories for bash scripts
Best use case
state-directory-manager is best used when you need a repeatable AI agent workflow instead of a one-off prompt.
Manage persistent state directories for bash scripts
Teams using state-directory-manager should expect a more consistent output, faster repeated execution, less prompt rewriting.
When to use this skill
- You want a reusable workflow that can be run more than once with consistent structure.
When not to use this skill
- You only need a quick one-off answer and do not need a reusable workflow.
- You cannot install or maintain the underlying files, dependencies, or repository context.
Installation
Claude Code / Cursor / Codex
Manual Installation
- Download SKILL.md from GitHub
- Place it in
.claude/skills/state-directory-manager/SKILL.mdinside your project - Restart your AI agent — it will auto-discover the skill
How state-directory-manager Compares
| Feature / Agent | state-directory-manager | Standard Approach |
|---|---|---|
| Platform Support | Not specified | Limited / Varies |
| Context Awareness | High | Baseline |
| Installation Complexity | Unknown | N/A |
Frequently Asked Questions
What does this skill do?
Manage persistent state directories for bash scripts
Where can I find the source code?
You can find the source code on GitHub using the link provided at the top of the page.
SKILL.md Source
# State Directory Manager
Patterns for managing persistent state, configuration, and cache directories in bash scripts following XDG Base Directory specification.
## When to Use This Skill
✅ **Use when:**
- Scripts need to persist data between runs
- Storing user preferences or configuration
- Caching results for performance
- Managing log files with rotation
- Creating portable CLI tools
❌ **Avoid when:**
- One-time scripts that don't need state
- Scripts that should be purely stateless
- When environment variables are sufficient
## Core Capabilities
### 1. XDG Base Directory Standard
Follow the XDG specification for directory locations:
```bash
#!/bin/bash
# ABOUTME: XDG Base Directory compliant state management
# ABOUTME: Cross-platform directory locations
# XDG Base Directories with fallbacks
XDG_CONFIG_HOME="${XDG_CONFIG_HOME:-$HOME/.config}"
XDG_DATA_HOME="${XDG_DATA_HOME:-$HOME/.local/share}"
XDG_STATE_HOME="${XDG_STATE_HOME:-$HOME/.local/state}"
XDG_CACHE_HOME="${XDG_CACHE_HOME:-$HOME/.cache}"
# Application-specific directories
APP_NAME="my-tool"
CONFIG_DIR="$XDG_CONFIG_HOME/$APP_NAME"
DATA_DIR="$XDG_DATA_HOME/$APP_NAME"
STATE_DIR="$XDG_STATE_HOME/$APP_NAME"
CACHE_DIR="$XDG_CACHE_HOME/$APP_NAME"
LOG_DIR="$STATE_DIR/logs"
# Initialize directories
init_directories() {
mkdir -p "$CONFIG_DIR"
mkdir -p "$DATA_DIR"
mkdir -p "$STATE_DIR"
mkdir -p "$CACHE_DIR"
mkdir -p "$LOG_DIR"
}
```
### 2. Workspace-Hub Pattern
Alternative using home directory (from workspace-hub scripts):
```bash
#!/bin/bash
# ABOUTME: Workspace-hub style state directory management
# ABOUTME: Simple $HOME/.app-name pattern
APP_NAME="workspace-hub"
APP_DIR="${HOME}/.${APP_NAME}"
# Directory structure
CONFIG_DIR="$APP_DIR/config"
DATA_DIR="$APP_DIR/data"
LOGS_DIR="$APP_DIR/logs"
CACHE_DIR="$APP_DIR/cache"
TEMP_DIR="$APP_DIR/tmp"
# Initialize with proper permissions
init_app_dirs() {
local dirs=("$CONFIG_DIR" "$DATA_DIR" "$LOGS_DIR" "$CACHE_DIR" "$TEMP_DIR")
for dir in "${dirs[@]}"; do
if [[ ! -d "$dir" ]]; then
mkdir -p "$dir"
chmod 700 "$dir" # Private by default
fi
done
}
# Clean old temp files
clean_temp() {
find "$TEMP_DIR" -type f -mtime +1 -delete 2>/dev/null || true
}
```
### 3. Configuration File Management
Read and write configuration files:
```bash
#!/bin/bash
# ABOUTME: Configuration file management
# ABOUTME: Key-value pairs with defaults
CONFIG_FILE="$CONFIG_DIR/config"
# Default configuration
declare -A DEFAULT_CONFIG=(
["parallel_workers"]="5"
["log_level"]="INFO"
["auto_sync"]="true"
["timeout"]="30"
)
# Initialize config with defaults
init_config() {
if [[ ! -f "$CONFIG_FILE" ]]; then
{
echo "# Configuration for $APP_NAME"
echo "# Generated: $(date)"
echo ""
for key in "${!DEFAULT_CONFIG[@]}"; do
echo "${key}=${DEFAULT_CONFIG[$key]}"
done
} > "$CONFIG_FILE"
fi
}
# Read config value
get_config() {
local key="$1"
local default="${2:-${DEFAULT_CONFIG[$key]:-}}"
if [[ -f "$CONFIG_FILE" ]]; then
local value
value=$(grep "^${key}=" "$CONFIG_FILE" 2>/dev/null | cut -d'=' -f2-)
echo "${value:-$default}"
else
echo "$default"
fi
}
# Write config value
set_config() {
local key="$1"
local value="$2"
init_config
if grep -q "^${key}=" "$CONFIG_FILE" 2>/dev/null; then
# Update existing
sed -i "s|^${key}=.*|${key}=${value}|" "$CONFIG_FILE"
else
# Add new
echo "${key}=${value}" >> "$CONFIG_FILE"
fi
}
# Load all config into associative array
load_config() {
declare -gA CONFIG
# Start with defaults
for key in "${!DEFAULT_CONFIG[@]}"; do
CONFIG[$key]="${DEFAULT_CONFIG[$key]}"
done
# Override with file values
if [[ -f "$CONFIG_FILE" ]]; then
while IFS='=' read -r key value; do
[[ "$key" =~ ^#.*$ || -z "$key" ]] && continue
CONFIG[$key]="$value"
done < "$CONFIG_FILE"
fi
}
# Usage
init_config
load_config
echo "Parallel workers: ${CONFIG[parallel_workers]}"
set_config "parallel_workers" "10"
```
### 4. State File Operations
Track persistent state between runs:
```bash
#!/bin/bash
# ABOUTME: State file operations
# ABOUTME: Track last run, progress, etc.
STATE_FILE="$STATE_DIR/state.json"
# Initialize state
init_state() {
if [[ ! -f "$STATE_FILE" ]]; then
cat > "$STATE_FILE" << EOF
{
"version": "1.0.0",
"created": "$(date -Iseconds)",
"last_run": null,
"run_count": 0,
"last_status": null
}
EOF
fi
}
# Get state value (requires jq)
get_state() {
local key="$1"
local default="${2:-null}"
if [[ -f "$STATE_FILE" ]] && command -v jq &>/dev/null; then
jq -r ".$key // $default" "$STATE_FILE"
else
echo "$default"
fi
}
# Update state value (requires jq)
set_state() {
local key="$1"
local value="$2"
init_state
if command -v jq &>/dev/null; then
local temp=$(mktemp)
jq ".$key = $value" "$STATE_FILE" > "$temp" && mv "$temp" "$STATE_FILE"
fi
}
# Record run
record_run() {
local status="$1"
set_state "last_run" "\"$(date -Iseconds)\""
set_state "last_status" "\"$status\""
set_state "run_count" "$(($(get_state run_count 0) + 1))"
}
# Simple key-value state (no jq required)
STATE_KV_FILE="$STATE_DIR/state.kv"
get_state_kv() {
local key="$1"
local default="$2"
if [[ -f "$STATE_KV_FILE" ]]; then
grep "^${key}=" "$STATE_KV_FILE" 2>/dev/null | cut -d'=' -f2- || echo "$default"
else
echo "$default"
fi
}
set_state_kv() {
local key="$1"
local value="$2"
mkdir -p "$(dirname "$STATE_KV_FILE")"
if [[ -f "$STATE_KV_FILE" ]] && grep -q "^${key}=" "$STATE_KV_FILE"; then
sed -i "s|^${key}=.*|${key}=${value}|" "$STATE_KV_FILE"
else
echo "${key}=${value}" >> "$STATE_KV_FILE"
fi
}
```
### 5. Cache Management
Implement caching with expiration:
```bash
#!/bin/bash
# ABOUTME: Cache management with TTL
# ABOUTME: Store and retrieve cached data
CACHE_TTL="${CACHE_TTL:-3600}" # 1 hour default
# Get cache file path
cache_path() {
local key="$1"
local hash=$(echo -n "$key" | md5sum | cut -c1-16)
echo "$CACHE_DIR/${hash}"
}
# Check if cache is valid
cache_valid() {
local key="$1"
local ttl="${2:-$CACHE_TTL}"
local path=$(cache_path "$key")
if [[ -f "$path" ]]; then
local age=$(($(date +%s) - $(stat -c %Y "$path" 2>/dev/null || stat -f %m "$path")))
[[ $age -lt $ttl ]]
else
return 1
fi
}
# Get from cache
cache_get() {
local key="$1"
local ttl="${2:-$CACHE_TTL}"
local path=$(cache_path "$key")
if cache_valid "$key" "$ttl"; then
cat "$path"
return 0
fi
return 1
}
# Set cache
cache_set() {
local key="$1"
local value="$2"
local path=$(cache_path "$key")
mkdir -p "$CACHE_DIR"
echo "$value" > "$path"
}
# Delete cache
cache_delete() {
local key="$1"
local path=$(cache_path "$key")
rm -f "$path"
}
# Clear all cache
cache_clear() {
rm -rf "$CACHE_DIR"/*
}
# Clean expired cache entries
cache_clean() {
local ttl="${1:-$CACHE_TTL}"
find "$CACHE_DIR" -type f -mmin "+$((ttl / 60))" -delete 2>/dev/null || true
}
# Usage with automatic caching
get_with_cache() {
local key="$1"
local command="$2"
local ttl="${3:-$CACHE_TTL}"
if cache_valid "$key" "$ttl"; then
cache_get "$key"
else
local result
result=$(eval "$command")
cache_set "$key" "$result"
echo "$result"
fi
}
# Example
result=$(get_with_cache "api_response" "curl -s https://api.example.com/data" 300)
```
### 6. Log File Management
Manage logs with rotation:
```bash
#!/bin/bash
# ABOUTME: Log file management with rotation
# ABOUTME: Automatic cleanup of old logs
LOG_FILE="$LOG_DIR/app.log"
LOG_MAX_SIZE=$((10 * 1024 * 1024)) # 10MB
LOG_MAX_FILES=5
# Initialize logging
init_logging() {
mkdir -p "$LOG_DIR"
touch "$LOG_FILE"
}
# Write to log
log_to_file() {
local level="$1"
shift
local message="$*"
local timestamp=$(date '+%Y-%m-%d %H:%M:%S')
echo "[$timestamp] $level: $message" >> "$LOG_FILE"
# Check if rotation needed
maybe_rotate_logs
}
# Rotate logs if needed
maybe_rotate_logs() {
if [[ -f "$LOG_FILE" ]]; then
local size=$(stat -c %s "$LOG_FILE" 2>/dev/null || stat -f %z "$LOG_FILE")
if [[ $size -gt $LOG_MAX_SIZE ]]; then
rotate_logs
fi
fi
}
# Perform log rotation
rotate_logs() {
# Remove oldest
rm -f "${LOG_FILE}.${LOG_MAX_FILES}"
# Shift existing
for ((i=LOG_MAX_FILES-1; i>=1; i--)); do
if [[ -f "${LOG_FILE}.$i" ]]; then
mv "${LOG_FILE}.$i" "${LOG_FILE}.$((i+1))"
fi
done
# Rotate current
if [[ -f "$LOG_FILE" ]]; then
mv "$LOG_FILE" "${LOG_FILE}.1"
touch "$LOG_FILE"
fi
}
# Clean old logs
clean_old_logs() {
local days="${1:-30}"
find "$LOG_DIR" -name "*.log*" -mtime "+$days" -delete 2>/dev/null || true
}
# View recent logs
tail_logs() {
local lines="${1:-50}"
tail -n "$lines" "$LOG_FILE"
}
# Search logs
search_logs() {
local pattern="$1"
grep -h "$pattern" "$LOG_DIR"/*.log* 2>/dev/null | tail -100
}
```
## Complete Example: State Manager Module
```bash
#!/bin/bash
# ABOUTME: Complete state directory manager
# ABOUTME: Reusable module for bash scripts
# ─────────────────────────────────────────────────────────────────
# State Directory Manager v1.0.0
# ─────────────────────────────────────────────────────────────────
# Application identity (override in your script)
: "${STATE_APP_NAME:=my-app}"
# Directory setup
STATE_BASE_DIR="${HOME}/.${STATE_APP_NAME}"
STATE_CONFIG_DIR="$STATE_BASE_DIR/config"
STATE_DATA_DIR="$STATE_BASE_DIR/data"
STATE_CACHE_DIR="$STATE_BASE_DIR/cache"
STATE_LOG_DIR="$STATE_BASE_DIR/logs"
STATE_TMP_DIR="$STATE_BASE_DIR/tmp"
# File paths
STATE_CONFIG_FILE="$STATE_CONFIG_DIR/config"
STATE_STATE_FILE="$STATE_DATA_DIR/state"
STATE_LOG_FILE="$STATE_LOG_DIR/app.log"
# Settings
STATE_CACHE_TTL="${STATE_CACHE_TTL:-3600}"
STATE_LOG_MAX_SIZE="${STATE_LOG_MAX_SIZE:-10485760}"
STATE_LOG_MAX_FILES="${STATE_LOG_MAX_FILES:-5}"
# ─────────────────────────────────────────────────────────────────
# Initialization
# ─────────────────────────────────────────────────────────────────
state_init() {
local dirs=(
"$STATE_CONFIG_DIR"
"$STATE_DATA_DIR"
"$STATE_CACHE_DIR"
"$STATE_LOG_DIR"
"$STATE_TMP_DIR"
)
for dir in "${dirs[@]}"; do
if [[ ! -d "$dir" ]]; then
mkdir -p "$dir"
chmod 700 "$dir"
fi
done
# Initialize files
[[ -f "$STATE_CONFIG_FILE" ]] || touch "$STATE_CONFIG_FILE"
[[ -f "$STATE_STATE_FILE" ]] || touch "$STATE_STATE_FILE"
[[ -f "$STATE_LOG_FILE" ]] || touch "$STATE_LOG_FILE"
}
# ─────────────────────────────────────────────────────────────────
# Config Functions
# ─────────────────────────────────────────────────────────────────
state_config_get() {
local key="$1"
local default="$2"
grep "^${key}=" "$STATE_CONFIG_FILE" 2>/dev/null | cut -d'=' -f2- || echo "$default"
}
state_config_set() {
local key="$1"
local value="$2"
if grep -q "^${key}=" "$STATE_CONFIG_FILE" 2>/dev/null; then
sed -i "s|^${key}=.*|${key}=${value}|" "$STATE_CONFIG_FILE"
else
echo "${key}=${value}" >> "$STATE_CONFIG_FILE"
fi
}
state_config_list() {
cat "$STATE_CONFIG_FILE" 2>/dev/null | grep -v '^#' | grep -v '^$'
}
# ─────────────────────────────────────────────────────────────────
# State Functions
# ─────────────────────────────────────────────────────────────────
state_get() {
local key="$1"
local default="$2"
grep "^${key}=" "$STATE_STATE_FILE" 2>/dev/null | cut -d'=' -f2- || echo "$default"
}
state_set() {
local key="$1"
local value="$2"
if grep -q "^${key}=" "$STATE_STATE_FILE" 2>/dev/null; then
sed -i "s|^${key}=.*|${key}=${value}|" "$STATE_STATE_FILE"
else
echo "${key}=${value}" >> "$STATE_STATE_FILE"
fi
}
# ─────────────────────────────────────────────────────────────────
# Cache Functions
# ─────────────────────────────────────────────────────────────────
state_cache_key() {
echo -n "$1" | md5sum | cut -c1-16
}
state_cache_get() {
local key="$1"
local ttl="${2:-$STATE_CACHE_TTL}"
local path="$STATE_CACHE_DIR/$(state_cache_key "$key")"
if [[ -f "$path" ]]; then
local age=$(($(date +%s) - $(stat -c %Y "$path" 2>/dev/null || stat -f %m "$path")))
if [[ $age -lt $ttl ]]; then
cat "$path"
return 0
fi
fi
return 1
}
state_cache_set() {
local key="$1"
local value="$2"
local path="$STATE_CACHE_DIR/$(state_cache_key "$key")"
echo "$value" > "$path"
}
state_cache_clear() {
rm -rf "$STATE_CACHE_DIR"/*
}
# ─────────────────────────────────────────────────────────────────
# Log Functions
# ─────────────────────────────────────────────────────────────────
state_log() {
local level="$1"
shift
local message="$*"
echo "[$(date '+%Y-%m-%d %H:%M:%S')] $level: $message" >> "$STATE_LOG_FILE"
# Auto-rotate
local size=$(stat -c %s "$STATE_LOG_FILE" 2>/dev/null || echo 0)
if [[ $size -gt $STATE_LOG_MAX_SIZE ]]; then
state_log_rotate
fi
}
state_log_rotate() {
rm -f "${STATE_LOG_FILE}.${STATE_LOG_MAX_FILES}"
for ((i=STATE_LOG_MAX_FILES-1; i>=1; i--)); do
[[ -f "${STATE_LOG_FILE}.$i" ]] && mv "${STATE_LOG_FILE}.$i" "${STATE_LOG_FILE}.$((i+1))"
done
mv "$STATE_LOG_FILE" "${STATE_LOG_FILE}.1"
touch "$STATE_LOG_FILE"
}
state_log_tail() {
tail -n "${1:-50}" "$STATE_LOG_FILE"
}
# ─────────────────────────────────────────────────────────────────
# Cleanup Functions
# ─────────────────────────────────────────────────────────────────
state_cleanup() {
# Clean temp files older than 1 day
find "$STATE_TMP_DIR" -type f -mtime +1 -delete 2>/dev/null || true
# Clean expired cache
find "$STATE_CACHE_DIR" -type f -mmin "+$((STATE_CACHE_TTL / 60))" -delete 2>/dev/null || true
# Clean old logs
find "$STATE_LOG_DIR" -name "*.log.*" -mtime +30 -delete 2>/dev/null || true
}
state_reset() {
rm -rf "$STATE_BASE_DIR"
state_init
}
# ─────────────────────────────────────────────────────────────────
# Auto-initialize
# ─────────────────────────────────────────────────────────────────
state_init
```
## Usage in Scripts
```bash
#!/bin/bash
# Your script that uses the state manager
# Set app name before sourcing
STATE_APP_NAME="my-tool"
# Source the state manager
source /path/to/state-manager.sh
# Now use it
state_config_set "api_key" "abc123"
api_key=$(state_config_get "api_key")
state_set "last_run" "$(date -Iseconds)"
state_log "INFO" "Script started"
# Use cache
if ! result=$(state_cache_get "api_response"); then
result=$(curl -s https://api.example.com/data)
state_cache_set "api_response" "$result"
fi
```
## Best Practices
1. **Use Standard Locations** - Follow XDG or `$HOME/.app-name`
2. **Initialize Early** - Call init before any operations
3. **Handle Permissions** - Use 700 for private data
4. **Clean Up Regularly** - Remove old temp/cache files
5. **Rotate Logs** - Prevent unbounded growth
## Resources
- [XDG Base Directory Spec](https://specifications.freedesktop.org/basedir-spec/basedir-spec-latest.html)
- [File Hierarchy Standard](https://refspecs.linuxfoundation.org/FHS_3.0/fhs/index.html)
---
## Version History
- **1.0.0** (2026-01-14): Initial release - extracted from workspace-hub patternsRelated Skills
llm-manager
Claude acts as manager/architect while delegating all coding to external LLM CLIs (Gemini, Codex, Qwen). Claude never writes code - only plans, delegates, and verifies. Use when user says "manage", "architect mode", "delegate to", or wants Claude to drive another LLM.
directory-naming-convention
Defines the directory naming convention.
data-pipeline-manager
Design and troubleshoot robust data pipelines with comprehensive quality validation, error handling, and monitoring capabilities for bioinformatics and data processing workflows
chromatin-state-inference
This skill should be used when users need to infer chromatin states from histone modification ChIP-seq data using chromHMM. It provides workflows for chromatin state segmentation, model training, state annotation.
agent-risk-manager
Expert risk manager specializing in comprehensive risk assessment, mitigation strategies, and compliance frameworks. Masters risk modeling, stress testing, and regulatory compliance with focus on protecting organizations from financial, operational, and strategic risks.
agent-project-manager
Expert project manager specializing in project planning, execution, and delivery. Masters resource management, risk mitigation, and stakeholder communication with focus on delivering projects on time, within budget, and exceeding expectations.
agent-manager-skill
Manage multiple local CLI agents via tmux sessions (start/stop/monitor/assign) with cron-friendly scheduling.
agent-customer-success-manager
Expert customer success manager specializing in customer retention, growth, and advocacy. Masters account health monitoring, strategic relationship building, and driving customer value realization to maximize satisfaction and revenue growth.
Active Directory Attacks
This skill should be used when the user asks to "attack Active Directory", "exploit AD", "Kerberoasting", "DCSync", "pass-the-hash", "BloodHound enumeration", "Golden Ticket", "Silver Ticket", "AS-REP roasting", "NTLM relay", or needs guidance on Windows domain penetration testing.
1k-state-management
Jotai state management patterns for OneKey. Use when working with atoms, global state, feature state, or context atoms. Triggers on jotai, atom, state, globalAtom, contextAtom, store, persistence, settings.
social-media-manager
Agente especialista em Social Media para múltiplas empresas (Multi-tenant). Cria estratégias semanais, gerencia perfis de marca e gera conteúdo (texto e imagem) em massa para publicação manual.
athlete-social-media-manager
Create brand-safe content for athletes. Personal branding strategy, engagement optimization, crisis communication, sponsor integration.