AI Execution Prompt

# AI Execution Prompt

Do not summarize this file.
Execute the workflow below.

You are operating inside an Agent Execution Template workspace.

This file only routes the workflow. Project workflows must be triggered by
slash commands. For ordinary questions, protocol explanations, design
discussion, or read-only consultation, read only the files needed to answer and
do not enter bootstrap / reconcile / strategy / execution. Do not write
`ai/project/task.md`, `ai/project/result.*`, or `ai/project/metrics.json`.

## Slash Commands

Official entrypoints:

- `/init`
- `/init-with-inbox`
- `/reconcile`
- `/strategy`
- `/apply-strategy`
- `/continue`
- `/next`
- `/doctor`
- `/update`
- `/refresh`
- `/improve-context`
- `/help`

If the user does not provide a slash command but explicitly asks to implement,
fix, modify, commit, publish, or execute project work, treat it as the
`/continue` execution entrypoint. Otherwise, answer read-only.

## Routing

After receiving a slash command, first read minimal state:

1. `ai/project/project.md` if present
2. `ai/project/task.md` if present
3. Shallow listings of `ai/project/inbox/`, `ai/project/inbox/ideas/`, and
   `ai/project/proposals/final-shape-updates/`

Then choose the mode:

- `/help`: show the slash command list and purposes read-only.
- `/next`: judge the next step read-only. Prioritize pending direction
  proposals, pending ideas, material waiting in inbox, missing project context,
  pending task confirmation, failed results, or the continue entrypoint. Do not
  modify files.
- `/doctor`: if the current task contract permits commands, run
  `node bin/agent-execution-template.js doctor` or the project-approved doctor
  command; otherwise tell the user to run
  `npx -y @wnlen/agent-execution-template doctor`.
- `/update`: if the current task contract explicitly permits protocol update,
  run update; otherwise tell the user to run
  `npx -y @wnlen/agent-execution-template update`.
- `/refresh` or `/improve-context`: if the current task contract explicitly
  permits context refresh, run refresh; otherwise tell the user to run
  `npx -y @wnlen/agent-execution-template refresh`.
- `/init`: follow `ai/template/bootstrap.md`, initialize project context, and
  stop after project-context confirmation.
- `/init-with-inbox`: inspect `ai/project/project.md` first. If it already
  exists and is not empty, placeholder-only, or clearly incomplete, follow
  `ai/template/reconcile.md` instead of bootstrapping again. If it is empty,
  placeholder-only, or clearly incomplete, follow `ai/template/bootstrap.md`
  and treat `ai/project/inbox/*.md` and `ai/project/inbox/raw/*.md` as part of
  the bootstrap input for this run; stop after project-context confirmation.
- `/reconcile`: follow `ai/template/reconcile.md`. Default to only
  `ai/project/inbox/*.md` and `ai/project/inbox/raw/*.md`; `processed/` does
  not trigger reconciliation, while `ideas/` should route to `/strategy` first.
- `/strategy`: read `ai/template/protocol.md`, `ai/template/rules/core.md`, and
  relevant direction refs, then follow `strategy_update`; produce the proposal
  and stop for human confirmation.
- `/apply-strategy`: read `ai/template/protocol.md` and
  `ai/template/rules/core.md`, and only apply a proposal the human has
  explicitly confirmed. If the proposal is still `proposed`, update it to
  `accepted` first.
- `/continue`: if `ai/project/project.md` is empty, placeholder-only, or
  incomplete, run `/init` first. If `ai/project/task.md` is empty,
  placeholder-only, or incomplete, draft it from the user's current goal and
  confirmed project context, then stop for human confirmation. Use Execution
  Mode only after `project.md` and `task.md` are ready enough to define
  identity, goal, scope, permission, and acceptance. Only then read
  `ai/template/protocol.md`, `ai/template/rules/core.md`, and
  `ai/template/execution-policy.md`.

## Task Draft Handoff

In Task Draft Mode:

1. Read confirmed `ai/project/project.md` and relevant `ai/project/refs/*.md`.
2. Infer goal, scope, acceptance, permissions, verification method, and initial
   risk from the user's current goal, project context, and repository facts; do
   not require the human to provide each field upfront.
3. Draft `ai/project/task.md`. For a single-L1, Green, low-risk task, default to
   a compact task contract: write only the goal, scope, acceptance, permissions,
   verification commands, and minimal `execution_policy.task_tree`. Use an
   expanded task contract only for multi-L1, Yellow/Red, cross-module,
   continuously executed, or highly uncertain tasks.
4. Before execution, list the L1 checklist, mark each L1 Green / Yellow / Red,
   and write it to `execution_policy.task_tree`. Use `normal` if there are
   fewer than 2 L1 tasks; automatically use `bounded_continuous` if there are 2
   or more L1 tasks. The complete default rules live in
   `ai/template/execution-policy.md`; do not mechanically copy internal control
   fields such as `checkpoint_budget` or `model_policy` into simple task drafts.
5. If this run creates or rewrites `ai/project/task.md`, set `readiness` to
   `draft_for_confirmation` and stop at the handoff; do not execute while the
   task is still a draft.
6. Enter Execution Mode only when an existing task is explicitly
   `ready_to_execute` and no Red preflight item exists; if it cannot execute,
   set it to `blocked`.
7. Do not modify source or business files in Task Draft Mode.

End Task Draft Mode with: