v0.2.0 · Available on npm and GitHub
Tool reference
Explore 26 tool groups and 228 registered actions in v0.2.0. Inspect their inputs and build a request for your MCP client.
Automatic workflow recording is unfinished, Playwright export does not generate runnable tests, and audits do not establish compliance or certification. Read compatibility and limits.
How to read this reference
Pass an explicit action in your tool arguments. This snapshot comes from the registered tool definitions in tools-spec.json.
Pass action in the tool arguments. Missing or unknown actions may resolve to a normalized match or the first action, with a warning.
Schemas preserve registered definitions, including defaults, enums, nested properties, and constraints. Defaults are documentation and are applied by individual handlers, not uniformly by the registry.
Metadata is taken from registered capabilities where available. Source-assessed risks/flags, source-observed artifacts and access mappings fill gaps and are explicitly labelled in each action.
Schematic tools/call payloads, without a JSON-RPC envelope. Replace <parameter> placeholders with values of the declared type and satisfy handler-specific requirements before use.
Output outlines are derived from handler source. They document known response fields without asserting runtime enforcement, required fields or exhaustive extension payloads.
Declared policy identifiers are FileForge flags, not Chrome permissions. For actions without registered capabilities, source access mappings describe possible access using those identifiers; they are not enforced permission declarations. Conditional access depends on input options and delegated steps.
Transport and common response format
{
"transport": {
"protocol": "MCP",
"serverTransport": "stdio",
"result": {
"content": [
{
"type": "text",
"text": "<JSON-serialized result object>"
}
]
},
"extensionBridge": {
"transport": "WebSocket",
"url": "ws://127.0.0.1:47821"
}
},
"responseFormats": {
"result": {
"description": "Common FileForge result fields. Fields vary by action; ok:false may also represent audit findings without an error object.",
"fields": {
"ok": {
"type": "boolean",
"description": "Operation or check outcome."
},
"data": {
"description": "Action-specific payload; may be any JSON value or absent when stored by reference."
},
"ref": {
"type": "string",
"description": "Optional reference to stored results."
},
"summary": {
"description": "Optional action-specific summary."
},
"warnings": {
"type": "array",
"items": {
"type": "string"
},
"description": "Optional execution or action-resolution warnings."
},
"meta": {
"type": "object",
"description": "Tool identity and optional action, URL, tab, timing, and resolution metadata."
},
"error": {
"type": "object",
"description": "Failure details, usually code and message, optionally details."
}
},
"successExample": {
"ok": true,
"data": {
"pong": true
},
"meta": {
"tool": "fileforge_runtime_action",
"action": "ping"
}
},
"errorExample": {
"ok": false,
"error": {
"code": "HANDLER_ERROR",
"message": "Illustrative failure",
"details": {}
},
"meta": {
"tool": "fileforge_runtime_action"
}
}
}
}
}26 of 26 tool groups · Open a group to inspect its actions.