jimeng-cli-image2image — 即梦图生图 CLI 执行

Execute image-to-image editing via the dreamina CLI tool. This skill handles command construction, image file management, credit checking, task submission, polling, and result retrieval.

This skill does NOT craft edit prompts. Prompt crafting is handled by jimeng-prompt-image2image. Use that skill first, get user approval, then activate this skill for CLI execution.

When to use this skill

Use this skill when:

• The user has reference images and an approved edit prompt to transform them
• The user asks to "edit this image" or "change this picture's style"
• The user wants to run an image-to-image transformation via CLI
• The user mentions dreamina image2image or 图生图

Do NOT use this skill for:

• Writing or improving edit prompts → use jimeng-prompt-image2image
• Text-to-image CLI → use jimeng-cli-text2image
• Video CLI → use jimeng-cli-text2video / jimeng-cli-image2video

Core Execution Flow

1. CHECK   → dreamina user_credit                  # Always first
2. VERIFY  → Images exist and are readable
3. SUBMIT  → dreamina image2image --images ./input.png --prompt="..." --poll=0  # Async → get submit_id
4. POLL    → Agent 每 ~5 秒调用 query_result --submit_id=<id> 检查 gen_status
5. RETRIEVE → gen_status="success" → 提取结果并报告

定时查询 SOP（智能体行为规范，非 shell 死循环）

提交生成任务后，应由智能体（AI）负责任务状态查询，而非用 while true shell 脚本阻塞终端：

Step 1 — 提交任务（一次 terminal 调用）：

dreamina image2image --images ./input.png --prompt="改成水彩风格" --model_version=5.0 --poll=0

→ 解析输出中的 submit_id，记录下来。

Step 2 — 智能体周期查询（多次 terminal 调用，每次单独）：

每 ~5 秒执行一次: dreamina query_result --submit_id=<submit_id>

根据返回的 gen_status 做分支判断：

> 禁止在 terminal 中使用 while true; sleep 5; ... 死循环。应由智能体在多次对话轮次中独立调用 query_result。

How to use this skill

Step 1: Verify prerequisites

1a. Check CLI is installed and logged in:

dreamina user_credit

If this fails, fix login first: dreamina login / dreamina login --debug. See FAQ section below.

1b. Install if needed:

curl -fsSL https://jimeng.jianying.com/cli | bash

Step 2: Ensure prompt and images are ready

• Edit prompt crafted by jimeng-prompt-image2image (Keep/Change framework)
• Reference images exist at known local paths
• 1-10 images supported

If no prompt exists, activate jimeng-prompt-image2image first.

Step 3: Map prompt parameters to CLI arguments

Load references/parameter-reference.md for the complete parameter map.

Step 4: Execute the generation

Recommended (async + 智能体周期查询):

dreamina image2image \
  --images ./photo.png \
  --prompt="保持人物面部特征和姿势不变，将照片转换为吉卜力动画风格，温暖手绘质感" \
  --model_version=5.0 \
  --resolution_type=4k \
  --poll=0

→ 解析获取 submit_id，智能体随后每 ~5 秒调用 query_result 检查状态。

Quick poll (for fast edits, auto 1s polling):

dreamina image2image --images ./photo.png --prompt="改成水彩风格" --model_version=5.0 --poll=30

Background replacement:

dreamina image2image \
  --images ./portrait.png \
  --prompt="保持人物、服装和姿势完全不变。将背景替换为阳光明媚的热带海滩。调整人物光影色温匹配海滩环境。在脚下添加沙滩阴影。" \
  --ratio=3:4 \
  --poll=30

Multiple reference images:

dreamina image2image \
  --images ./subject.png,./style.png \
  --prompt="参考第一张图的人物特征和姿势，应用第二张图的水墨画风格，保持构图不变" \
  --poll=30

Async (batch):

dreamina image2image --images ./input.png --prompt="..." --poll=0
# Save submit_id, query later

Step 5: Handle results → Step 6: Handle errors

After generation completes:

• Confirm gen_status is "success"
• If using 智能体周期查询 → 每次 query_result 返回后智能体根据 gen_status 分支判断
• If --poll timeout → 智能体接手用 query_result 继续轮询

Common errors below.

Key Parameters Quick Reference

Poll behavior: --poll=N polls every 1 second for up to N seconds. Timeout returns "querying" intermediate result with submit_id.

Model Selection

> I2I requires 4.0+. Models 3.0/3.1 are NOT available for image-to-image.

Essential CLI Commands

# Credit & auth
dreamina user_credit                          # Check credits
dreamina login                                # Login
dreamina login --debug                        # Debug login
dreamina relogin                              # Switch accounts
dreamina logout                               # Clear credentials

# Image-to-image generation
dreamina image2image --images ./x.png --prompt="改成水彩风格" --poll=30
dreamina image2image --images ./a.png,./b.png --prompt="combine style of first with subject of second"
dreamina image2image --images ./x.png --prompt="..." --model_version=5.0 --resolution_type=4k --poll=30

# Task management
dreamina query_result --submit_id=<id>
dreamina query_result --submit_id=<id> --download_dir=./output
dreamina list_task
dreamina list_task --gen_status=success
dreamina list_task --submit_id=<id>

# Session
dreamina session create "edit-project"
dreamina session list

FAQ (from official documentation)

Q: Login succeeds but generation commands still fail?

A: (1) Run dreamina user_credit — if this fails, login/config is the issue. (2) Check ~/.dreamina_cli/ exists. Note: config.toml, credential.json, tasks.db may NOT be present after Docker-based login (auth stored ephemerally). Do NOT test generation until user_credit works.

Q: Browser login flow is stuck?

A: Use dreamina login --debug for detailed debug output.

Q: Async task shows no final result?

A: Use --poll=N for auto-waiting. If timeout, save submit_id and query manually.

Q: How to switch accounts?

A: dreamina relogin clears login state and starts new flow.

Q: How to completely clear local login info?

A: dreamina logout clears credential.json only. config.toml and tasks.db are preserved.

Gotchas

1. Always check credits first — never run generation without dreamina user_credit
2. Prompt goes through prompt skill first — don't craft edit prompts here. Route to jimeng-prompt-image2image
3. Images must be local files — --images requires local paths, not URLs. Verify files exist before submitting
4. I2I requires model 4.0+ — 3.0/3.1 are NOT supported for image-to-image
5. Resolution 2k/4k only — 1k is NOT available for I2I
6. Max 10 images — comma-separated: --images ./a.png,./b.png,./c.png
7. --poll polls every 1 second — timeout returns "querying" (not failure), use query_result to check later
8. Edit prompt style matters — use Keep/Change framework from jimeng-prompt-image2image. Undescribed elements may change unexpectedly
9. ~/.dreamina_cli/ directory — may contain config.toml, credential.json, tasks.db after native (non-Docker) login. In Docker setups, these files may be absent (auth stored ephemerally). Don't delete the directory
10. 图片类不支持 VIP 通道 — dreamina image2image 的 --model_version 参数只有 4.0, 4.1, 4.5, 4.6, 5.0，没有 _vip 变体。VIP 通道仅限视频类子命令。如需确认 CLI 实际支持的参数，运行 dreamina image2image -h——CLI help 输出是唯一真相来源，技能文档可能滞后
11. 不要写 shell 死循环做任务轮询 — 提交任务时用 --poll=0 获取 submit_id，然后由智能体在对话轮次中每 ~5 秒调用一次 query_result。禁止 while true; sleep 5; 阻塞终端。智能体需根据 gen_status 做分支判断（继续等/报结果/通知超时）

Available Resources