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 agentmods add skills/vinsonconsulting/limner/captioned-graphicnpx skills add vinsonconsulting/limner --skill captioned-graphicgit clone --depth 1 https://github.com/vinsonconsulting/limnerWrote 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/vinsonconsulting/limner/captioned-graphic)<a href="https://agentmods.dev/skills/vinsonconsulting/limner/captioned-graphic"><img src="https://agentmods.dev/badge/skills/vinsonconsulting/limner/captioned-graphic.svg" alt="Measured on agentmods" 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 | $0.00045 | $0.01053 |
| Opus 5 | $0.00023 | $0.00526 |
| Sonnet 5 | $0.00009 | $0.00211 |
| Haiku 4.5 | $0.00005 | $0.00105 |
Grade A, and why
captioned-graphic 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 5d 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 — 80 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Captioned graphic
Limner is an independent third-party project built on Anthropic's CMA platform; it is not an Anthropic or Claude product.
Use this skill to set words on an image: a headline, a caption bar, a title slide, or a quote card. Compose's renderText op turns a small JSX-shaped layout into a PNG, and the watermark op composites it over the base image. The text is exact and legible because you set it, unlike text a generator bakes in. The reference at the end gives the op details.
Procedure
- Write the text. Keep it short: a headline is one line, a caption a brief phrase.
- Choose the layout. Decide where the text sits (a bottom caption band, a centered title) and the type style (size, color, weight).
- Render the text. Use the renderText op with a JSX-shaped layout (flexbox style), the canvas size, and the IBM Plex Sans font. Give the canvas a transparent background, or a translucent band behind the words for legibility over a busy image.
- Composite over the image. Use the watermark op to place the rendered text onto the base image: at (0, 0) if you rendered at full size, or at an offset for a strip.
- Deliver one asset. Convert to JPEG or WebP if file size matters, then return the captioned image through its capability URL.
Judgment
- Render the text at the base image size and composite at (0, 0) for pixel-exact placement; render a smaller strip to position a caption like a stamp.
- Put a translucent band behind text over a busy or light image so it stays readable.
- Keep one type style across a set so the captions read as a series.
- The built-in font is IBM Plex Sans. Do not assume other families are available.
Reference
The table below is generated from the Limner guidance core (@limner/core), the
same source the MCP captioned-graphic prompt serves, so this skill and that
prompt cannot drift. Do not edit the generated region by hand; run
pnpm --filter @limner/limner-agent gen:skills instead.
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.
- 5d ago First seen · 80 lines · 45 tokens per session scan A dbab14a4ad0e
captioned-graphic is a skill published in the GitHub repository vinsonconsulting/limner (1 stars, last pushed 23d ago), licensed Apache-2.0. It adds 45 tokens to every session and 1,053 once invoked, about $0.0002 per session on Opus 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 2026-08-31.
Other skills, from other repositories
flux-3-product-ads
Use when building a finished product ad from FLUX 3 - shot design, voiceover, action-to-word sync, evidence-gated copy, deterministic assembly, and QC gates that catch clipped audio, floating products, off-model plates, and reports that claim a pass the build did not give.
bfl-api
BFL FLUX API integration guide covering endpoints, async polling patterns, rate limiting, error handling, webhooks, and regional endpoints with Python and TypeScript code examples.
flux-3-archival-formats
Use when a FLUX 3 video needs a period or archival look. Covers naming a recording format instead of a mood, per-format artifacts, and templates.
flux-3-prompt-doctor
Use when diagnosing a FLUX 3 brief before generation. Resolve missing, conflicting, or schema-changing requirements.
flux-image-best-practices
Comprehensive guide for BFL FLUX image generation models. Covers prompting, T2I, I2I, structured JSON, hex colors, typography, multi-reference editing, and model-specific best practices for FLUX.2 and FLUX.1 families.
flux-3-keyframes-continuation
Use when a FLUX 3 video must be built from supplied images or video. Covers keyframes (i2v) and continuation (v2v).