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/vanillagreencom/kendex/size-ratchetnpx skills add vanillagreencom/kendex --skill size-ratchetgit clone --depth 1 https://github.com/vanillagreencom/kendexWrote 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/vanillagreencom/kendex/size-ratchet)<a href="https://agentmods.dev/skills/vanillagreencom/kendex/size-ratchet"><img src="https://agentmods.dev/badge/skills/vanillagreencom/kendex/size-ratchet.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 | $0.00029 | $0.01369 |
| Opus 5 | $0.00015 | $0.00685 |
| Sonnet 5 | $0.00006 | $0.00274 |
| Haiku 4.5 | $0.00003 | $0.00137 |
Grade C, and why
size-ratchet scanned grade C with 1 finding 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.
Hidden instructionshighPrompt injection
Directives inside HTML comments, invisible characters or bidirectional overrides are read by the model and not by the person reviewing the file.
<!-- kendex:project-instructions:start --> How it starts
The opening of the file, as written. The whole thing — 67 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Project Instructions
Problems with a kendex-owned skill go through kendex report; check ownership in the file first.
Size Ratchet
No tracked file gets bigger than its threshold, and files already over it only shrink. Existing offenders are frozen in a baseline at their current sizes; everything else stays at or under its path class's threshold. Markdown is measured in bytes and code in lines. Baseline rows only go down or away while their unit stays the same. Reference and exception rules are README.md § Trusted HEAD baseline.
.agents/skills/size-ratchet/scripts/size-ratchet # check (pre-PR / CI)
.agents/skills/size-ratchet/scripts/size-ratchet --staged # check what a commit records (git hook)
.agents/skills/size-ratchet/scripts/size-ratchet --update # tighten the baseline
.agents/skills/size-ratchet/scripts/size-ratchet --seed # write the FIRST baseline
Flags, SIZE_RATCHET_* keys, and exit codes: size-ratchet --help. Verdict classes, semantics, units, the shipped class list, baseline/excludes formats, and seeding: README.md. Collection and maintenance notes: DEVELOPMENT.md.
Responding to a failure
Every size diagnostic names the file, its size, the baseline row it violated, the deciding threshold (class pattern or default), and the remedy: split at a concept seam.
Reference, repoint, raise, and unit-change behavior is one rule: README.md § Trusted HEAD baseline. A shrunk row under --staged is already lowered and staged by the run itself.
A NUL is refused, not measured. Any tracked blob carrying a NUL in its leading 8000 bytes is what git calls binary, whatever unit its class counts in and whether or not the size exclusion list covers it. The gate exits 2 naming the path and the byte's offset, in every mode, --seed and --update included. Write the escape the language spells the byte with; a real binary asset is declared binary or -diff in .gitattributes, the only exemption.
What ships with it
16 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.
- DEVELOPMENT.md 14 KB
- README.md 8.9 KB
- references/threshold-change.md 2.4 KB
- scripts/lib/settings.sh 22 KB runs code
- scripts/size-ratchet 52 KB
- tests/baseline-relocated.test.sh 11 KB runs code
- tests/binary-content.test.sh 16 KB runs code
- tests/collection-fail-closed.test.sh 22 KB runs code
- tests/fail-closed-routing.test.sh 17 KB runs code
- tests/path-classes.test.sh 20 KB runs code
- tests/ratchet-directions.test.sh 26 KB runs code
- tests/seed.test.sh 5.8 KB runs code
- tests/settings-and-config.test.sh 23 KB runs code
- tests/shipped-defaults.test.sh 36 KB runs code
- tests/staged-scope.test.sh 20 KB runs code
- tests/update-tightens.test.sh 14 KB runs code
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 Changed · -50 lines 44e5f05f3e36
- 2d ago Changed · +11 lines scan A → C b4ce9ef2005f
- 3d ago Changed · -8 lines a8b20163fa0b
- 6d ago First seen · 114 lines · 29 tokens per session scan A 9a11239b017d
size-ratchet is a skill published in the GitHub repository vanillagreencom/kendex (74 stars, last pushed today), licensed MIT. It adds 29 tokens to every session and 1,369 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it C with 1 finding (hidden instructions). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
surf
Control Chrome browser via CLI for testing, automation, and debugging. Use when the user needs browser automation, screenshots, form filling, page inspection, network/CPU emulation, DevTools streaming, or AI queries via ChatGPT/Gemini/Perplexity/Grok/AI Studio.
prose-review
Review prose written for others (e.g., user-facing documentation, prompts for other LLMs, reports, plans, inline comments, docstrings) for local jargon leakage, orphaned references, missing grounding, and audience or genre mismatch. Always run on prose you produce or materially edit, except for routine conversational…
text-search
Search indexed text corpora with qmd. For indexed content, prefer qmd over grep.
xcodebuildmcp
Build/test Xcode projects via the XcodeBuildMCP MCP server using a local CLI wrapper for pi (no MCP support). Use when the user mentions Xcode, xcodebuild, iOS builds/tests, or XcodeBuildMCP.
agent-browser
Automates browser interactions for web testing, form filling, screenshots, and data extraction. Use when the user needs to navigate websites, interact with web pages, fill forms, take screenshots, test web applications, or extract information from web pages.
deep-x-research
Deep, exhaustive research on a topic across X (Twitter) by driving Grok (x.com/i/grok) through surf. Use when the user wants comprehensive X research on a concept, technique, trend, tool, or creator scene; needs categorized findings with every claim traceable to post URLs; or when a single Grok query is not enough.