Keep a record of each run's tool calls so bots can't claim work they never did
The wrap-up turn now gets the run's actual tool calls, each bot message keeps them as bot_messages.tool_log, and history follows recorded bot messages with an [App record] line. The coordinator also gets update_bot in workspaces, so it can fill a bot's core prompt from there. Co-Authored-By: Claude Opus 5.5 <[email protected]>
This commit is contained in:
@@ -152,4 +152,49 @@ describe("botMessagesToHarness with mentions", () => {
|
||||
{ role: "assistant", content: "[bot:3] is on it" },
|
||||
]);
|
||||
});
|
||||
|
||||
test("follows a recorded bot message with the app's record of its tools", () => {
|
||||
expect(
|
||||
botMessagesToHarness({
|
||||
messages: [
|
||||
message("user", "fill in its prompt"),
|
||||
{
|
||||
...message("bot", "Done."),
|
||||
tool_log: JSON.stringify([
|
||||
{ name: "fetch_url", ok: true },
|
||||
{ name: "fetch_url", ok: false },
|
||||
]),
|
||||
},
|
||||
message("user", "and the other one?"),
|
||||
{ ...message("bot", "Saved too."), tool_log: "[]" },
|
||||
],
|
||||
}),
|
||||
).toEqual([
|
||||
{ role: "user", content: "fill in its prompt" },
|
||||
{
|
||||
role: "assistant",
|
||||
content: "Done.\n\n[App record: this message's run called fetch_url x2 (1 failed).]",
|
||||
},
|
||||
{ role: "user", content: "and the other one?" },
|
||||
{
|
||||
role: "assistant",
|
||||
content: "Saved too.\n\n[App record: this message's run called no tools.]",
|
||||
},
|
||||
]);
|
||||
});
|
||||
|
||||
test("adds no record for messages whose run wasn't recorded", () => {
|
||||
expect(
|
||||
botMessagesToHarness({
|
||||
messages: [
|
||||
message("user", "hi"),
|
||||
{ ...message("bot", "hello"), tool_log: undefined },
|
||||
{ ...message("bot", "again"), tool_log: "not json" },
|
||||
],
|
||||
}),
|
||||
).toEqual([
|
||||
{ role: "user", content: "hi" },
|
||||
{ role: "assistant", content: "hello\n\nagain" },
|
||||
]);
|
||||
});
|
||||
});
|
||||
|
||||
Reference in New Issue
Block a user