Services Inventory¶

Runtime Services / Processes¶

Service	Entry Point	Type	Port	Description
amplihack CLI	`src/amplihack/cli.py:main`	CLI Process	-	Main entry point; parses subcommands and dispatches to launchers
Recipe Runner	`src/amplihack/recipe_cli/recipe_command.py`	CLI Process	-	Executes YAML recipe files with step orchestration
Fleet CLI	`src/amplihack/fleet/_cli_commands.py`	CLI Process (Click)	-	Multi-agent fleet orchestration via tmux panes
Auto Mode	`src/amplihack/launcher/auto_mode.py`	CLI Process	-	Autonomous agentic execution with configurable max turns
Docker Manager	`src/amplihack/docker/manager.py`	Container Launcher	-	Spawns amplihack sessions inside Docker containers
Amplifier Launcher	`src/amplihack/launcher/amplifier.py`	CLI Process	-	Launches Microsoft Amplifier agent backend
Copilot Launcher	`src/amplihack/launcher/copilot.py`	CLI Process	-	Launches GitHub Copilot CLI agent backend
Codex Launcher	`src/amplihack/launcher/codex.py`	CLI Process	-	Launches OpenAI Codex CLI agent backend
Claude Launcher	`src/amplihack/launcher/core.py`	CLI Process	-	Launches Claude Code with plugin injection and proxy setup
Rust Trial	`src/amplihack/rust_trial.py`	CLI Process	-	Experimental Rust-based amplihack binary launcher
Trace Logger	`src/amplihack/tracing/trace_logger.py`	Background Writer	-	JSONL trace logging for LLM call observability
Plugin Manager	`src/amplihack/plugin_manager/`	Library	-	Install/uninstall/verify plugins at ~/.amplihack/.claude/

Non-Runtime Components¶

Component	Path	Type	Description
Skills	`.claude/skills/`	Markdown Definitions	Claude Code skill definitions auto-discovered at session start
Agents	`.claude/agents/amplihack/`	Markdown Definitions	Specialized agent personas (architect, builder, reviewer, etc.)
Hooks	`.claude/tools/amplihack/hooks/`	Python Scripts	Session lifecycle hooks (start, stop, pre/post tool use)
Recipes	`.claude/recipes/`	YAML Files	Multi-step workflow definitions executed by recipe runner
Workflows	`.claude/workflows/`	Markdown Files	Development workflow blueprints (DEFAULT_WORKFLOW, etc.)
Commands	`.claude/commands/`	Markdown/YAML	Slash command definitions (/dev, /analyze, /improve, /fix)
Context Files	`.claude/context/`	Markdown Files	Project philosophy, patterns, trust guidelines
Scenarios	`.claude/scenarios/`	Mixed	Production-ready user-facing tool implementations
XPIA Defense	`src/amplihack/security/`	Python Package	Cross-Prompt Injection Attack detection and defense
Bundle Generator	`src/amplihack/bundle_generator/`	Python Package	Generates amplifier bundles from agent specifications
Knowledge Builder	`src/amplihack/knowledge_builder/`	Python Package	Builds knowledge graphs from codebase analysis
Blarify (Vendored)	`src/amplihack/vendor/blarify/`	Python Package	Code intelligence indexing via tree-sitter and LSP
Memory System	`src/amplihack/memory/`	Python Package	Persistent memory with Kuzu graph DB backend
Power Steering	`src/amplihack/power_steering/`	Python Package	Context-adaptive prompt augmentation
Adaptive Context	`src/amplihack/context/adaptive/`	Python Package	Environment detection and context strategy injection
Settings Generator	`src/amplihack/settings_generator/`	Python Package	Generates Claude Code settings.json with hook configuration
UVX Manager	`src/amplihack/uvx/`	Python Package	UVX packaging and environment management
Meta Delegation	`src/amplihack/meta_delegation/`	Python Package	Cross-platform agent delegation (Claude/Copilot/Codex/Amplifier)
LSP Detector	`src/amplihack/lsp_detector/`	Python Package	Language Server Protocol availability detection
Mode Detector	`src/amplihack/mode_detector/`	Python Package	Detects local vs plugin vs UVX installation mode
Eval Framework	`src/amplihack/eval/`	Python Package	Agent evaluation and benchmarking
Safety Module	`src/amplihack/safety/`	Python Package	Safety guardrails for agentic execution