fix(todo): address todo_write review feedback

This commit is contained in:
Tianyi Cui
2026-06-29 19:44:38 +08:00
parent 55088d1cfc
commit 9f1caf7c5b
8 changed files with 36 additions and 29 deletions

View File

@@ -146,6 +146,7 @@ forever:
session('assistant/message' {content, usage?}) log records what tool dispatch uses
each tool-call (sequential, abort-checked between calls):
session('tool/call'); ctx.tools.execute() ⟵ waterfall tools/execute
tool execution may append tool-owned session events, e.g. `todo/write`
session('tool/result')
drain steering → session('steering/message'); emit agent/steering
emit agent/step-end

View File

@@ -4,7 +4,7 @@ Status: implemented
## Problem
The harness gives the model bash and subagent tools but no way to record a structured task list. A todo list serves two co-equal purposes: it steers the model to plan multi-step work and keep exactly one task active (anti-drift on long tasks), and it gives the human a live progress checklist. The ACP protocol has a native `plan` sessionUpdate that editors (Zed) already render, but the bridge never emitted one. Every reference coding agent surveyed (claude-code, opencode, codex, oh-my-pi, pi) ships some form of this; the harness had nothing.
The harness gives the model bash and subagent tools but no way to record a structured task list. A todo list serves two co-equal purposes: it steers the model to plan multi-step work and keep the active task unambiguous (at most one active, exactly one while work remains), and it gives the human a live progress checklist. The ACP protocol has a native `plan` sessionUpdate that editors (Zed) already render, but the bridge never emitted one. Every reference coding agent surveyed (claude-code, opencode, codex, oh-my-pi, pi) ships some form of this; the harness had nothing.
## Decision