Paste into Claude Code / CLAUDE.md task

# DevAgentRadar → Claude Code

You are helping me adopt a real coding-assistant change. Work only from the facts below. Do not invent features.

## Context
Assistant: Cline
Proposal: Register workspace-scoped scheduled tasks and durable todo agendas in Cline
Summary: Cline SDK v0.0.76 lets agents create scheduled tasks and a durable todo agenda scoped to the workspace, generate images mid-turn with history persistence, and load skill slash commands via the skills tool so transcripts stay clean.
Primary source: https://github.com/cline/cline/releases/tag/sdk/sdk/v0.0.76

## Why it matters
This release changes how long-running agent work and skills should be structured. Schedules and todos are now first-class and workspace-scoped, so recurring or multi-step work should be registered as agenda items instead of ad-hoc chat prompts. Model-driven image generation is persisted in session history and exports, so visual outputs can be part of normal turns. Skill slash commands load through the skills tool (transcript keeps `/my-skill ...` only; instructions arrive once), which means skill usage and logging should assume tool-based loading, not pasted SKILL.md bodies. Hook fixes (PreToolUse contextModification as `<hook_context>`, awaited PostToolUse with cancel/contextModification) and live detached command output further reward explicit hook and command patterns over workarounds.

## Suggested actions
1. In each active Cline workspace, instruct the agent: "Create a durable todo agenda for this repo and register workspace-scoped scheduled tasks for our recurring jobs (tests, dependency refresh, docs sync); list the registered schedules and confirm they are scoped to this workspace only."
2. Update team skill docs: "Invoke skills only via slash commands (e.g. `/my-skill ...`) so they load through the skills tool; do not paste SKILL.md into the user message. Verify the transcript shows the slash invocation, not the full skill body, and that skill instructions arrive once."
3. For design or UI tasks, prompt: "When the selected model supports image generation, generate images during the turn and confirm they appear in session history and exports; do not rely on external image tools for those models."
4. Revise PreToolUse/PostToolUse hooks to rely on SDK v0.0.76 behavior: deliver `contextModification` for the next engine turn as a `<hook_context>` block (tool name + call id), await PostToolUse (within 120s), and honor `contextModification` and `cancel` instead of fire-and-forget.
5. For shell automation on Windows, rewrite PowerShell agent commands assuming `$ErrorActionPreference='Stop'` fail-fast, and use `run_commands` with the structured `{ command }` form (no `args` key) so spaced commands route through the shell without ENOENT.

## Config surfaces this release may change
- skills — check your repo before applying
- hooks — check your repo before applying


## Your job
1. Restate the change in one sentence.
2. Propose a minimal plan for my repo (or a throwaway pilot).
3. Implement only what I approve; prefer small diffs and tests.
4. Call out risks (permissions, breaking APIs, cost).

Start by confirming you understood the proposal.