# Startup Artifacts

## Required Reads

- `AGENTS.md`
- `.ai/environment/tools.ai.yaml`
- `ai-plan/public/README.md`
- the selected `ai-plan/public/<topic>/todos/` directories
- the selected `ai-plan/public/<topic>/traces/` directories

## AI-Plan Selection Heuristics

- Match the current branch or worktree against `ai-plan/public/README.md` first.
- If the index maps the current worktree to topics, inspect those topics in listed order before scanning anything else.
- Match the user's wording against public todo and trace file names next.
- Prefer the newest matching trace when several candidates describe the same feature area.
- If one file records a clearer recovery point than a newer but vague file, prefer the clearer recovery point.
- Ignore `ai-plan/public/archive/**` unless the user explicitly requests historical recovery context.
- Even inside an active topic, prefer the root `todos/` and `traces/` entry files first; only read `archive/` when the
  active files point there or when the user asks for historical detail.
- If a matching `ai-plan/private/<branch-or-worktree>/` directory exists, use it only as private context for the current worktree.

## Complexity Defaults

- `simple`: keep everything local, no subagent
- `medium`: keep design local, optionally use one `explorer` for parallel read-only discovery
- `complex`: keep architecture and integration local, delegate only bounded non-blocking subtasks

## Model Defaults

- `explorer`: `gpt-5.1-codex-mini`
- `worker`: `gpt-5.4`

## Startup Summary Template

Use a short update before execution:

`Read AGENTS.md, the environment inventory, ai-plan/public/README.md, and the relevant public ai-plan artifacts. This looks like a <task-state> <complexity> task. I will <delegate-or-not> and start with <first-step>.`