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 decocms/studio --skill tool-scriptinggit clone --depth 1 https://github.com/decocms/studioWrote 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/decocms/studio/tool-scripting)<a href="https://agentmods.dev/skills/decocms/studio/tool-scripting"><img src="https://agentmods.dev/badge/skills/decocms/studio/tool-scripting/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/decocms/studio/tool-scripting"><img src="https://agentmods.dev/badge/skills/decocms/studio/tool-scripting.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00078 | $0.00925 |
| Opus 5.5 | $0.00031 | $0.00370 |
| Sonnet 5.5 | $0.00016 | $0.00185 |
| Haiku 4.5 | $0.00008 | $0.00093 |
Grade A, and why
tool-scripting 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 — 108 lines — stays where its author put it; the contents beside it link to each section on GitHub.
tool-scripting — call org tools from scripts
The workspace carries a materialized catalog of the org's tools:
.deco/tools/<TOOL>.json— one JSON Schema per tool:{ name, description, inputSchema, outputSchema }.deco/tools/.endpoint.json— the run's pre-authenticated MCP endpoint. ThetypegenCLI and client discover it automatically (walking up from cwd) — no flags, keys, or env needed. The daemon refreshes it; a reconnect picks up new credentials.
Discover tools (cheap)
Browse the catalog from disk instead of loading every schema into context:
ls .deco/tools/
cat .deco/tools/SEND_EMAIL.json
or typegen tools (list names + descriptions) / typegen tools SEND_EMAIL
(one tool's full schema).
One-off call from the shell
typegen call SEND_EMAIL '{"to":"[email protected]","subject":"hi"}'
Prints the tool's structured output as JSON; non-zero exit and a message on
stderr on failure. If typegen is not on PATH, use
bunx @decocms/typegen call ....
Bulk calls — write a script
For N calls (a tool per row/file/item), never loop through your own tool-calling — write a script, run it, read back the summary. Results belong in a file, not in your context.
Set up a scratch project so the user's repo stays untouched. Cloud sandboxes carry the typegen package built from the same Studio revision; Desktop falls back to the published package. Then run from the repo root (endpoint discovery walks up from cwd):
mkdir -p /tmp/toolrun && cd /tmp/toolrun
if [ -f /opt/sandbox-daemon/typegen.tgz ]; then
bun add /opt/sandbox-daemon/typegen.tgz
else
bun add @decocms/typegen
fi
cd - && bun run /tmp/toolrun/bulk.ts
// /tmp/toolrun/bulk.ts
import { createStudioClient } from "@decocms/typegen";
const client = createStudioClient(); // discovers .deco/tools/.endpoint.json
const rows: { email: string }[] = JSON.parse(
await Bun.file("input.json").text(),
);
const out = Bun.file("results.jsonl").writer();
let ok = 0;
let failed = 0;
for (const [i, row] of rows.entries()) {
try {
const result = await client.SEND_EMAIL({ to: row.email, subject: "hi" });
out.write(`${JSON.stringify({ i, ok: true, result })}\n`);
ok++;
} catch (err) {
out.write(`${JSON.stringify({ i, ok: false, error: String(err) })}\n`);
failed++;
}
if (i % 25 === 0) console.error(`progress ${i}/${rows.length}`);
}
out.end();
await client.close();
console.log(JSON.stringify({ ok, failed, total: rows.length }));
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 First seen · 108 lines · 78 tokens per session scan A 8f68d3f2fc94
tool-scripting is a skill published in the GitHub repository decocms/studio (409 stars, last pushed today), licensed MIT. It adds 78 tokens to every session and 925 once invoked, about $0.0003 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.
Other skills, from other repositories
ai-agent-evaluation-benchmarking
Autonomous AI agent benchmark evaluation register: task completion rates, planning accuracy, tool invocation precision, and cost benchmarks.
otto-pr-review-loop
Run Polly and OCR on a PR until all findings settle.
deploy-docker-compose
Run the Omnigent server as a Docker compose stack (server + Postgres) on any Docker host — your laptop, a VPS, EC2 by hand, or as the base layer of any container-platform deploy. Invoke when the user wants to build the image, bring up the compose stack, debug the stack on a host they already have, or extend the stack…
security-audit
Audit a codebase or directory for security issues (hardcoded secrets, injection, unsafe deserialization, weak crypto, authz gaps) and produce a structured findings report. Use when the user asks for a security review, an audit, or to check code for vulnerabilities. Report only — never fix.
Workspace Data Analyst
Analyze CSV files in the workspace and summarize insights.
map-explain
Produce a natural-language explanation of a specific Cortex function, business capability, or schema element from the committed knowledge graph.