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.
git clone --depth 1 https://github.com/coltonbearden/carrelWrote 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/commands/coltonbearden/carrel/color)<a href="https://agentmods.dev/commands/coltonbearden/carrel/color"><img src="https://agentmods.dev/badge/commands/coltonbearden/carrel/color.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 5.1 | $0.00044 | $0.00917 |
| Opus 5 | $0.00022 | $0.00458 |
| Sonnet 5 | $0.00009 | $0.00183 |
| Haiku 4.5 | $0.00004 | $0.00092 |
Grade A, and why
color 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 yesterday.
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.
What it actually says
Handle this color request: $ARGUMENTS
Run the carrel CLI via Bash. carrel color is a group with palette, check and convert; map the user's request onto the real subcommands and flags in the --help blocks below (regenerated from the CLI by scripts/sync_plugins.py; if the installed carrel color --help differs, trust the installed version — never invent flags):
Usage: carrel color [OPTIONS] COMMAND [ARGS]...
Color tools: dominant palette, ICC profile conversion, WCAG contrast.
Options:
--json Machine-readable JSON output.
--help Show this message and exit.
Commands:
check WCAG contrast ratio of FG on BG (hex colors, e.g.
convert Convert SRC into an ICC profile and embed the profile in the output.
palette Dominant colors of SRC as hex + proportion (median-cut quantization).
Usage: carrel color check [OPTIONS] FG BG
WCAG contrast ratio of FG on BG (hex colors, e.g. #333 #fafafa).
Options:
--json Machine-readable JSON output.
--help Show this message and exit.
Usage: carrel color convert [OPTIONS] SRC
Convert SRC into an ICC profile and embed the profile in the output.
CMYK targets are written as JPEG/TIFF (PNG cannot store CMYK).
Options:
--to-profile P Target ICC profile: .icc path or builtin alias (cmyk, gray, p3, srgb). [required]
-o, --out FILE Output path [default: <SRC>.<profile>.png/.jpg].
--json Machine-readable JSON output.
--help Show this message and exit.
Usage: carrel color palette [OPTIONS] SRC
Dominant colors of SRC as hex + proportion (median-cut quantization).
Human mode shows rich color swatches; --json prints one JSON array of {"hex", "proportion"} sorted
by coverage.
Options:
--n INTEGER RANGE Number of colors to extract. [default: 8; 1<=x<=256]
--json Machine-readable JSON output.
--help Show this message and exit.
- palette SRC (
--n, default 8): dominant colors as[{hex, proportion}]sorted by coverage (median-cut). Good for "what are this image's brand colors" — present them as hex codes with rough percentages, largest first. - check FG BG: WCAG contrast ratio of two hex colors (
#333 #fafafa). Report the ratio and whether it passes AA (4.5:1 normal text, 3:1 large text) and AAA (7:1); when it fails, propose a darker/lighter variant and re-check it. - convert SRC --to-profile P: converts into an ICC profile (
.iccpath or aliascmyk,gray,p3,srgb) and embeds it. CMYK output is written as JPEG/TIFF because PNG cannot store CMYK. Output defaults to<SRC>.<profile>.png/.jpg; the original is never modified. To preview a profile without converting, usecarrel proofinstead.
Use --json for every subcommand and interpret the result conversationally. Exit code 4 means the input is not an image or a color could not be parsed — ask for a valid hex value.
Requires the carrel CLI on PATH. If carrel is not found, tell the user to install it with uv tool install carrel (see the repo's INSTALL notes), or run it as uv run carrel ... from the carrel repo root.
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.
- yesterday First seen · 74 lines · 44 tokens per session scan A dafd5cafac0d
color is a command published in the GitHub repository coltonbearden/carrel (1 stars, last pushed 2d ago), licensed MIT. It adds 44 tokens to every session and 917 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 2026-09-05.
Other commands, from other repositories
frankx-infogenius
Premium-First Image Generation - Research → Brand-Aligned Style → High-Quality Output.
content-visual-brief
Visual brief command. Detailed design instructions, color palettes, and AI image prompts for social visuals, banners, and video frames.
brand-new
Create a new visual brand profile via guided Q&A, optionally bootstrapping colors and fonts from an existing website, logo, or brand-guidelines PDF.
checklist
Generate a custom checklist for the current feature based on user requirements.
clarify
Identify underspecified areas in the current feature spec by asking up to 5 highly targeted clarification questions and encoding answers back into the spec.
specify
Create or update the feature specification from a natural language feature description.