Executor Usage

What it does

When your agent uses it

• Working with third-party services
• MCP/OpenAPI/GraphQL tools
• Auth-managed actions
• Remote data that should go through Executor instead of direct shell scripts

Example prompts

• “/executor-usage”

Workflow steps

5 steps, taken from the first numbered list in SKILL.md.

1. 1Search for the tool.
2. 2Pick the best path from the search results.
3. 3Describe the tool to get compact TypeScript shapes.
4. 4Call the tool using the full namespace path.
5. 5Normalize the result before returning it when the tool returns MCP-style content blocks.

What it can do on your machine

Read from SKILL.md and the folder at commit b744f0c. It shows what the files ask for, not the result of running them.

• Tool permissions

  Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

  From allowed-tools in the SKILL.md frontmatter.

• Runs code

  No scripts in the folder and no shell commands in SKILL.md (its code samples are typescript).

  From the folder's file list and the shell code blocks in SKILL.md.

• Network

  No URLs in SKILL.md.

  From URLs in SKILL.md, links to its own repository left out.

• Credentials

  Names no API keys, tokens, secrets or passwords.

  From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Executor Usage loads about 1.4k tokens when it runs. Until then it costs about 137 tokens; SKILL.md has 553 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it

~137

When it runs · the whole SKILL.md, loaded when a task matches

~1.4k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from jeremyosih/pi-executor at commit b744f0c, republished under its MIT licence (© jeremyosih). 553 words, ~1,443 tokens.

name

executor-usage

description

Load this skill before using the `execute` tool. It teaches the Executor sandbox calling model: discover tools with `tools.search({ query, limit })`, inspect them with `tools.describe.tool({ path })`, inspect configured sources with `tools.executor.sources.list({})`, then call the real tool as `tools.<namespace>.<tool>(args)`. Use when working with third-party services, SaaS APIs, MCP/OpenAPI/GraphQL tools, auth-managed actions, or remote data that should go through Executor instead of direct shell scripts or ad hoc HTTP code.

metadata.short-description

Required calling pattern for Executor `execute`

const matches = await tools.search({ query: "linear issues", limit: 5 });
const path = matches[0]?.path;
if (!path) return "No matching tools found.";

const details = await tools.describe.tool({ path });
console.log(details.inputTypeScript);
console.log(details.outputTypeScript);

const result = await tools.mcp_linear_app.list_issues({
  project: "<project-id>",
  limit: 5,
});

return result;

{
  content: [{ type: "text", text: "{...json...}" }],
  structuredContent?: {...}
}

const unwrap = (value: any) => {
  if (value?.structuredContent) return value.structuredContent;

  const text = value?.content?.find?.((item: any) => item?.type === "text")?.text;
  if (typeof text !== "string") return value;

  try {
    return JSON.parse(text);
  } catch {
    return value;
  }
};

const raw = await tools.mcp_linear_app.list_projects({ query: "gitinspect", limit: 10 });
return unwrap(raw);

await tools.search({ namespace: "mcp_linear_app", query: "issues", limit: 10 });

© jeremyosih, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

Just SKILL.md in skills/executor-usage of jeremyosih/pi-executor.

Open the folder on GitHubat commit b744f0c