Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add Bliveren/CrossGen --skill crossgen-artistgit clone --depth 1 https://github.com/Bliveren/CrossGenWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/bliveren/crossgen/crossgen-artist)<a href="https://agentmods.dev/skills/bliveren/crossgen/crossgen-artist"><img src="https://agentmods.dev/badge/skills/bliveren/crossgen/crossgen-artist/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/bliveren/crossgen/crossgen-artist"><img src="https://agentmods.dev/badge/skills/bliveren/crossgen/crossgen-artist.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00041 | $0.02327 |
| Opus 5.5 | $0.00016 | $0.00931 |
| Sonnet 5.5 | $0.00008 | $0.00465 |
| Haiku 4.5 | $0.00004 | $0.00233 |
Grade A, and why
crossgen-artist scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 104 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CrossGen Artist
Use this skill when an agent needs to create or modify images through a local CrossGen installation. CrossGen owns provider credentials, model capability checks, the durable queue, and Gallery state; do not bypass it by calling an image provider directly unless the user explicitly asks for a separate integration.
Choose a transport
- Prefer the installed CrossGen MCP server when
crossgen_*tools are available. The model-facing names are usuallymcp__crossgen__crossgen_<tool>. - Otherwise use the JSON CLI (
crossgen ... --json). Discover the executable withcrossgen doctor --agent --json; do not assume an app path or disclose API keys. - If the host is DeepSeek Harness, read references/harness-mcp.md for the stdio bridge configuration. Read references/mcp-tools.md for the MCP contract and references/cli-fallback.md for equivalent CLI calls.
Standard workflow
- Check readiness. Call
crossgen_config_statusorcrossgen doctor --agent --json, thencrossgen_provider_listandcrossgen_models_list. Treat the model catalog as authoritative for provider/model compatibility. - Clarify the visual task. Convert the user's intent into a concrete prompt: subject, composition, style, lighting, materials, text requirements, aspect ratio, and exclusions. Preserve user-provided wording; only add structure that improves reproducibility. See references/model-workflows.md.
- Select the operation. Use generation for prompt-only output, edit for one or more reference images, and inpaint only when a mask is available and the selected adapter advertises support. Pass
providerId/modelonly when needed; otherwise use the active compatible configuration. - Confirm immediately before side effects. Generation, editing, cancellation, retry, Gallery mutations, path disclosure, and export require explicit user authorization. For MCP calls this is
confirm: true; for CLI calls use--yes. Never infer confirmation from an earlier conversational turn when the request has materially changed. - Use idempotency for paid retries. Set a stable
idempotencyKeywhen a request may be retried or a host may reconnect. Do not submit a second paid request merely because the first call returned before completion. - Monitor the durable job. Use
waitMsonly as a short convenience wait. Then pollcrossgen_job_status(orcrossgen job status) until a terminal state, with a bounded backoff. On failure, report the diagnostic and retry only with authorization. - Handle results through Gallery. Use
crossgen_gallery_listandcrossgen_asset_inspectfor metadata. Export withcrossgen_asset_exportonly when the user specifies a destination and confirms it; do not rely on undisclosed absolute paths. - Report precisely. Include the model/provider used, job id, terminal status, asset id(s), and any next action. Do not include API keys or local absolute paths unless explicitly requested and confirmed.
What ships with it
5 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 1mo ago Changed · +3 lines 008b94c69ed8
- 1mo ago Changed · +13 lines 70f1b6ae9743
- 1mo ago Changed · +29 lines b6ac0dcc9eee
- 2mo ago Changed · +20 lines · +2 tokens per session 18d0db895c83
- 2mo ago First seen · 39 lines · 39 tokens per session scan A 760b23c92d11
crossgen-artist is a skill published in the GitHub repository Bliveren/CrossGen (12 stars, last pushed 28d ago), licensed MIT. It adds 41 tokens to every session and 2,327 once invoked, about $0.0002 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
crossgen-artist compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| crossgen-artistBliveren/CrossGen | Skill | 12 | 28d ago | MIT | A | 41 | Codex |
| ima2lidge-jun/ima2-gen | Skill | 782 | 1mo ago | MIT | A | 25 | Claude Code, Codex |
| ima2-frontlidge-jun/ima2-gen | Skill | 782 | 1mo ago | MIT | A | 149 | Claude Code, Codex |
| ima2-uiuxlidge-jun/ima2-gen | Skill | 782 | 1mo ago | MIT | A | 118 | Claude Code, Codex |
| electron-expertpersonamanagmentlayer/pcl | Skill | 44 | 11d ago | Apache-2.0 | A | 59 | Claude Code |
| local-image-genDandreYang/local-image-gen | Skill | 61 | 2mo ago | MIT | C | 126 | Codex |
| ocxlidge-jun/opencodex | Skill | 18k | 2d ago | MIT | A | 59 | Claude Code, Codex |