lexicons and xrpc
lexicon
computer.socialweb.cicd.process.read
socialweb-guest-agent answers a job's interleaved stdout and stderr from a byte cursor, waiting up to waitMs for more. The agent keeps the last 16 MiB of each job's output. Served only on the agent's shell unix socket (/run/socialweb-guest-agent/shell.sock, root 0600), with no auth beyond the socket; the same tool the agent's /mcp server offers as read.
input
| field | type | description |
|---|---|---|
id * | string | The job id from start. |
cursor | integer | The byte offset to read from: 0, or the cursor of the previous answer. default 0 minimum 0 |
maxBytes | integer | The most bytes to answer; at most 262144, 65536 when absent. default 65536 minimum 1 maximum 262144 |
waitMs | integer | When there is nothing new and the job runs, how long to wait for output, in milliseconds (at most 30000). minimum 0 maximum 30000 |
output
| field | type | description |
|---|---|---|
output * | string | The output from the cursor on, never ending inside a UTF-8 character. |
cursor * | integer | The cursor for the next read. |
done * | boolean | True once the job exited and the cursor reached the end of its output. |
exitCode | integer | The job's exit status, once it exited. |
dropped | integer | How many bytes before the kept output the cursor asked for and the agent no longer holds. |
errors
| error | description |
|---|---|
InvalidRequest | 405: not a POST; 413: the body is over 8 MiB; 400: the arguments are not valid or there is no job with that id. |
used in
no page of these docs names it yet.
raw schema: lexicons/computer/socialweb/cicd/process/read.json
{
"lexicon": 1,
"id": "computer.socialweb.cicd.process.read",
"defs": {
"main": {
"type": "procedure",
"description": "socialweb-guest-agent answers a job's interleaved stdout and stderr from a byte cursor, waiting up to waitMs for more. The agent keeps the last 16 MiB of each job's output. Served only on the agent's shell unix socket (/run/socialweb-guest-agent/shell.sock, root 0600), with no auth beyond the socket; the same tool the agent's /mcp server offers as read.",
"input": {
"encoding": "application/json",
"schema": {
"type": "object",
"required": [
"id"
],
"properties": {
"id": {
"type": "string",
"description": "The job id from start."
},
"cursor": {
"type": "integer",
"minimum": 0,
"default": 0,
"description": "The byte offset to read from: 0, or the cursor of the previous answer."
},
"maxBytes": {
"type": "integer",
"minimum": 1,
"maximum": 262144,
"default": 65536,
"description": "The most bytes to answer; at most 262144, 65536 when absent."
},
"waitMs": {
"type": "integer",
"minimum": 0,
"maximum": 30000,
"description": "When there is nothing new and the job runs, how long to wait for output, in milliseconds (at most 30000)."
}
}
}
},
"output": {
"encoding": "application/json",
"schema": {
"type": "object",
"required": [
"output",
"cursor",
"done"
],
"properties": {
"output": {
"type": "string",
"description": "The output from the cursor on, never ending inside a UTF-8 character."
},
"cursor": {
"type": "integer",
"description": "The cursor for the next read."
},
"done": {
"type": "boolean",
"description": "True once the job exited and the cursor reached the end of its output."
},
"exitCode": {
"type": "integer",
"description": "The job's exit status, once it exited."
},
"dropped": {
"type": "integer",
"description": "How many bytes before the kept output the cursor asked for and the agent no longer holds."
}
}
}
},
"errors": [
{
"name": "InvalidRequest",
"description": "405: not a POST; 413: the body is over 8 MiB; 400: the arguments are not valid or there is no job with that id."
}
]
}
}
}