autopilot

Autopilot

Development workflow skills for Claude Code — your agent's operating system.
14 skills that turn Claude from a general-purpose agent into a disciplined developer.
Install once, works across all projects. Zero configuration required.

English  |  正體中文

---

The Problem

AI coding agents start from zero every session:

• A task needs 5 modules changed — the agent dives in without a plan and forgets half
• You ask "should we use X or Y?" — the agent picks one without researching trade-offs
• The agent fixes a tricky bug — next session, hits the exact same issue and doesn't remember
• Session context compresses — hard-won solutions and architectural decisions disappear

The Solution

Autopilot gives your agent standard operating procedures that enforce discipline:

Skill	What It Does	Without It
dev-flow	Evaluates task size, routes to commit or plan+project; includes session lifecycle (start/end) and goal alignment	Agent dives in without a plan
survey	Dual-agent research (researcher + skeptic)	Agent picks first option it finds
think-tank	6-role debate for strategic decisions	Single-perspective analysis
ceo-agent	Autonomous execution with CEO-level judgment	Agent asks permission for everything
quality-pipeline	Unified quality gate: test → scan → review	Inconsistent quality checks
project-lifecycle	Plan → bootstrap → structure → archive	Projects left unfinished or unarchived
learn	Auto-records knowledge from failures; includes knowledge health audit	Same mistakes repeated across sessions
retro	Engineering retrospective from git history	No visibility into work patterns
next	Scan all work sources, recommend highest-priority task; includes improvement processing	No visibility into what to do next
team	Multi-agent parallelization with dependency analysis	Solo execution when parallelism would help
profiling	Evidence-first performance profiling methodology	Agent guesses from code instead of measuring
test-strategy	Test pyramid, baseline management, feature flag levels	Inconsistent test coverage, no strategy
audit	Systematic comparison between implementations	Differences missed in manual review
debug	Evidence-first debugging: logs, tools, then code	Agent guesses instead of investigating

---

Key Features

Three Modes of Operation

Autopilot provides three distinct cognitive modes for different situations:

dev-flow — Guided Development (default)

The entry point for all development tasks. Evaluates task size and routes accordingly:

You: "Add WebSocket compression"

Claude (with dev-flow):
  1. Size assessment: L (crosses network + protocol + client modules)
  2. → Creates plan, project directory, feature branch
  3. → Implements phase by phase, quality gate each phase
  4. → Archives project on completion

Claude (without dev-flow):
  → Starts grep-ing the codebase immediately
  → No plan, no phases, no quality gates

dev-flow also handles the session lifecycle — health checks on startup, knowledge review, goal alignment on context continuation. You don't invoke these separately; dev-flow absorbs them.

ceo-agent — Autonomous Execution

When you want outcomes, not involvement. The agent becomes CEO; you become the Board.

You: "CEO mode. Handle the reconnect system. Level 3, you decide everything."

CEO startup:
  1. OKR — concrete success criteria (not vague "make it work")
  2. Involvement level — how often to report (every step / phase / just results)
  3. Scope mode — Expand / Selective / Hold / Reduce
  4. No-go zones — what's absolutely off-limits

Then: autonomous execution within DOA (Delegation of Authority)

The CEO agent applies 10 cognitive patterns from great CEOs (Bezos's two-way doors, Munger's inversion reflex, Jobs's focus as subtraction) and follows the Boil the Lake principle — AI makes completeness nearly free, so always choose the complete implementation over shortcuts.

CEO cannot self-audit. Like corporate governance, quality-pipeline and code-review run independently.

think-tank — Multi-Perspective Debate