讓 Agent 與 JifChat 合作，解鎖 AI 生成潛力Generate images & videos in your agent with JifChat

You are about to give your user the ability to generate and refine images and videos through JifChat — every result lands in their JifChat infinite canvas, editable anytime.

1. Install the JifChat skill. Run:
npx skills add JifChat/skills

On Hermes, `hermes skills install JifChat/skills` also works. If neither works, copy the `jifchat/` folder from https://github.com/JifChat/skills into `~/.hermes/skills/jifchat` (keep `SKILL.md` and `references/` together). From now on, apply the skill whenever the user asks to generate, edit, or refine images, video, or audio.

2. Connect the user's JifChat account. If you do not already hold a `JIFCHAT_API_KEY`, ask the user for their JifChat API key (`jc_...`) — they can find it in JifChat under Settings → Balance → API Keys (shown once at creation). Store it per your harness's secret convention (environment variable or key store). Never echo the key into chat or write it to any file the user didn't ask for.

3. Verify and report ready. Call `GET {{ORIGIN}}/api/v1/canvas/projects` with the key — a 200 means connected (report the project count; skip credit info unless `GET /api/v1/canvas/balance` also exists). From now on, any generation request routes through JifChat following the skill: results are delivered in chat together with a canvas link the user can open anytime to edit, re-run, or chain into video.

You are about to give your user the ability to generate and refine images and videos through JifChat — every result lands in their JifChat infinite canvas, editable anytime.

1. Install the JifChat skill. Run:
npx skills add JifChat/skills

If that fails, copy the `jifchat/` folder from https://github.com/JifChat/skills into `~/.claude/skills/jifchat` (keep `SKILL.md` and `references/` together). From now on, apply the skill whenever the user asks to generate, edit, or refine images, video, or audio.

2. Connect the user's JifChat account. If you do not already hold a `JIFCHAT_API_KEY`, ask the user for their JifChat API key (`jc_...`) — they can find it in JifChat under Settings → Balance → API Keys (shown once at creation). Store it per your harness's secret convention (environment variable or key store). Never echo the key into chat or write it to any file the user didn't ask for.

3. Verify and report ready. Call `GET {{ORIGIN}}/api/v1/canvas/projects` with the key — a 200 means connected (report the project count; skip credit info unless `GET /api/v1/canvas/balance` also exists). From now on, any generation request routes through JifChat following the skill: results are delivered in chat together with a canvas link the user can open anytime to edit, re-run, or chain into video.

You are about to give your user the ability to generate and refine images and videos through JifChat — every result lands in their JifChat infinite canvas, editable anytime.

1. Install the JifChat skill. Run:
npx skills add JifChat/skills

If that fails, copy the `jifchat/` folder from https://github.com/JifChat/skills into `~/.grok/skills/jifchat` (keep `SKILL.md` and `references/` together). From now on, apply the skill whenever the user asks to generate, edit, or refine images, video, or audio.

2. Connect the user's JifChat account. If you do not already hold a `JIFCHAT_API_KEY`, ask the user for their JifChat API key (`jc_...`) — they can find it in JifChat under Settings → Balance → API Keys (shown once at creation). Store it per your harness's secret convention (environment variable or key store). Never echo the key into chat or write it to any file the user didn't ask for.

3. Verify and report ready. Call `GET {{ORIGIN}}/api/v1/canvas/projects` with the key — a 200 means connected (report the project count; skip credit info unless `GET /api/v1/canvas/balance` also exists). From now on, any generation request routes through JifChat following the skill: results are delivered in chat together with a canvas link the user can open anytime to edit, re-run, or chain into video.