feat: artificial messages 1

This commit is contained in:
jif-oai
2026-02-03 16:31:37 +00:00
parent 88598b9402
commit 0b2b8c9012
14 changed files with 442 additions and 60 deletions
+1
View File
@@ -1940,6 +1940,7 @@ dependencies = [
"strum_macros 0.27.2",
"sys-locale",
"tempfile",
"thiserror 2.0.18",
"tracing",
"ts-rs",
"uuid",
+9 -9
View File
@@ -201,6 +201,7 @@ use crate::util::backoff;
use crate::windows_sandbox::WindowsSandboxLevelExt;
use codex_async_utils::OrCancelExt;
use codex_otel::OtelManager;
use codex_protocol::artificial_messages::ArtificialMessage;
use codex_protocol::config_types::CollaborationMode;
use codex_protocol::config_types::Personality;
use codex_protocol::config_types::ReasoningSummary as ReasoningSummaryConfig;
@@ -1802,15 +1803,10 @@ impl Session {
self.services
.otel_manager
.counter("codex.model_warning", 1, &[]);
let item = ResponseItem::Message {
id: None,
role: "user".to_string(),
content: vec![ContentItem::InputText {
text: format!("Warning: {}", message.into()),
}],
end_turn: None,
phase: None,
let item = ArtificialMessage::ModelWarning {
body: message.into(),
};
.to_response_item();
self.record_conversation_items(ctx, &[item]).await;
}
@@ -5573,10 +5569,14 @@ mod tests {
match last {
ResponseItem::Message { role, content, .. } => {
assert_eq!(role, "user");
let expected_warning = ArtificialMessage::ModelWarning {
body: "too many unified exec processes".to_string(),
}
.render();
assert_eq!(
content,
&vec![ContentItem::InputText {
text: "Warning: too many unified exec processes".to_string(),
text: expected_warning,
}]
);
}
@@ -2,6 +2,7 @@ use super::*;
use crate::truncate;
use crate::truncate::TruncationPolicy;
use codex_git::GhostCommit;
use codex_protocol::artificial_messages::ArtificialMessage;
use codex_protocol::models::ContentItem;
use codex_protocol::models::FunctionCallOutputContentItem;
use codex_protocol::models::FunctionCallOutputPayload;
@@ -60,6 +61,15 @@ fn user_input_text_msg(text: &str) -> ResponseItem {
}
}
fn demo_skill_message() -> String {
ArtificialMessage::Skill {
name: "demo".to_string(),
path: "skills/demo/SKILL.md".to_string(),
body: "body".to_string(),
}
.render()
}
fn function_call_output(call_id: &str, content: &str) -> ResponseItem {
ResponseItem::FunctionCallOutput {
call_id: call_id.to_string(),
@@ -445,9 +455,7 @@ fn drop_last_n_user_turns_ignores_session_prefix_user_messages() {
user_input_text_msg(
"# AGENTS.md instructions for test_directory\n\n<INSTRUCTIONS>\ntest_text\n</INSTRUCTIONS>",
),
user_input_text_msg(
"<skill>\n<name>demo</name>\n<path>skills/demo/SKILL.md</path>\nbody\n</skill>",
),
user_input_text_msg(&demo_skill_message()),
user_input_text_msg("<user_shell_command>echo 42</user_shell_command>"),
user_input_text_msg("turn 1 user"),
assistant_msg("turn 1 assistant"),
@@ -464,9 +472,7 @@ fn drop_last_n_user_turns_ignores_session_prefix_user_messages() {
user_input_text_msg(
"# AGENTS.md instructions for test_directory\n\n<INSTRUCTIONS>\ntest_text\n</INSTRUCTIONS>",
),
user_input_text_msg(
"<skill>\n<name>demo</name>\n<path>skills/demo/SKILL.md</path>\nbody\n</skill>",
),
user_input_text_msg(&demo_skill_message()),
user_input_text_msg("<user_shell_command>echo 42</user_shell_command>"),
user_input_text_msg("turn 1 user"),
assistant_msg("turn 1 assistant"),
@@ -480,9 +486,7 @@ fn drop_last_n_user_turns_ignores_session_prefix_user_messages() {
user_input_text_msg(
"# AGENTS.md instructions for test_directory\n\n<INSTRUCTIONS>\ntest_text\n</INSTRUCTIONS>",
),
user_input_text_msg(
"<skill>\n<name>demo</name>\n<path>skills/demo/SKILL.md</path>\nbody\n</skill>",
),
user_input_text_msg(&demo_skill_message()),
user_input_text_msg("<user_shell_command>echo 42</user_shell_command>"),
];
@@ -492,9 +496,7 @@ fn drop_last_n_user_turns_ignores_session_prefix_user_messages() {
user_input_text_msg(
"# AGENTS.md instructions for test_directory\n\n<INSTRUCTIONS>\ntest_text\n</INSTRUCTIONS>",
),
user_input_text_msg(
"<skill>\n<name>demo</name>\n<path>skills/demo/SKILL.md</path>\nbody\n</skill>",
),
user_input_text_msg(&demo_skill_message()),
user_input_text_msg("<user_shell_command>echo 42</user_shell_command>"),
user_input_text_msg("turn 1 user"),
assistant_msg("turn 1 assistant"),
@@ -510,9 +512,7 @@ fn drop_last_n_user_turns_ignores_session_prefix_user_messages() {
user_input_text_msg(
"# AGENTS.md instructions for test_directory\n\n<INSTRUCTIONS>\ntest_text\n</INSTRUCTIONS>",
),
user_input_text_msg(
"<skill>\n<name>demo</name>\n<path>skills/demo/SKILL.md</path>\nbody\n</skill>",
),
user_input_text_msg(&demo_skill_message()),
user_input_text_msg("<user_shell_command>echo 42</user_shell_command>"),
user_input_text_msg("turn 1 user"),
assistant_msg("turn 1 assistant"),
+8 -2
View File
@@ -146,6 +146,7 @@ pub fn parse_turn_item(item: &ResponseItem) -> Option<TurnItem> {
#[cfg(test)]
mod tests {
use super::parse_turn_item;
use codex_protocol::artificial_messages::ArtificialMessage;
use codex_protocol::items::AgentMessageContent;
use codex_protocol::items::TurnItem;
use codex_protocol::items::WebSearchItem;
@@ -284,6 +285,12 @@ mod tests {
#[test]
fn skips_user_instructions_and_env() {
let skill_message = ArtificialMessage::Skill {
name: "demo".to_string(),
path: "skills/demo/SKILL.md".to_string(),
body: "body".to_string(),
}
.render();
let items = vec![
ResponseItem::Message {
id: None,
@@ -316,8 +323,7 @@ mod tests {
id: None,
role: "user".to_string(),
content: vec![ContentItem::InputText {
text: "<skill>\n<name>demo</name>\n<path>skills/demo/SKILL.md</path>\nbody\n</skill>"
.to_string(),
text: skill_message,
}],
end_turn: None,
phase: None,
@@ -1,3 +1,4 @@
use codex_protocol::artificial_messages::ArtificialMessage;
use serde::Deserialize;
use serde::Serialize;
@@ -64,18 +65,12 @@ impl SkillInstructions {
impl From<SkillInstructions> for ResponseItem {
fn from(si: SkillInstructions) -> Self {
ResponseItem::Message {
id: None,
role: "user".to_string(),
content: vec![ContentItem::InputText {
text: format!(
"<skill>\n<name>{}</name>\n<path>{}</path>\n{}\n</skill>",
si.name, si.path, si.contents
),
}],
end_turn: None,
phase: None,
ArtificialMessage::Skill {
name: si.name,
path: si.path,
body: si.contents,
}
.to_response_item()
}
}
@@ -84,6 +79,15 @@ mod tests {
use super::*;
use pretty_assertions::assert_eq;
fn demo_skill_message() -> String {
ArtificialMessage::Skill {
name: "demo-skill".to_string(),
path: "skills/demo/SKILL.md".to_string(),
body: "body".to_string(),
}
.render()
}
#[test]
fn test_user_instructions() {
let user_instructions = UserInstructions {
@@ -146,18 +150,15 @@ mod tests {
panic!("expected one InputText content item");
};
assert_eq!(
text,
"<skill>\n<name>demo-skill</name>\n<path>skills/demo/SKILL.md</path>\nbody\n</skill>",
);
let expected = demo_skill_message();
assert_eq!(text, &expected);
}
#[test]
fn test_is_skill_instructions() {
assert!(SkillInstructions::is_skill_instructions(&[
ContentItem::InputText {
text: "<skill>\n<name>demo-skill</name>\n<path>skills/demo/SKILL.md</path>\nbody\n</skill>"
.to_string(),
text: demo_skill_message(),
}
]));
assert!(!SkillInstructions::is_skill_instructions(&[
+8 -1
View File
@@ -23,6 +23,7 @@ use codex_core::protocol::Op;
use codex_core::protocol::SessionSource;
use codex_otel::OtelManager;
use codex_protocol::ThreadId;
use codex_protocol::artificial_messages::ArtificialMessage;
use codex_protocol::config_types::CollaborationMode;
use codex_protocol::config_types::ModeKind;
use codex_protocol::config_types::ReasoningSummary;
@@ -343,7 +344,13 @@ async fn resume_includes_initial_messages_and_sends_prior_items() {
);
let pos_permissions = messages
.iter()
.position(|(role, text)| role == "developer" && text.contains("<permissions instructions>"))
.position(|(role, text)| {
role == "developer"
&& matches!(
ArtificialMessage::parse(text),
Ok(ArtificialMessage::Permission { .. })
)
})
.expect("permissions message");
let pos_user_instructions = messages
.iter()
@@ -3,6 +3,7 @@ use codex_core::protocol::COLLABORATION_MODE_CLOSE_TAG;
use codex_core::protocol::COLLABORATION_MODE_OPEN_TAG;
use codex_core::protocol::EventMsg;
use codex_core::protocol::Op;
use codex_protocol::artificial_messages::ArtificialMessage;
use codex_protocol::config_types::CollaborationMode;
use codex_protocol::config_types::ModeKind;
use codex_protocol::config_types::Settings;
@@ -90,7 +91,10 @@ async fn no_collaboration_instructions_by_default() -> Result<()> {
let input = req.single_request().input();
let dev_texts = developer_texts(&input);
assert_eq!(dev_texts.len(), 1);
assert!(dev_texts[0].contains("<permissions instructions>"));
assert!(matches!(
ArtificialMessage::parse(&dev_texts[0]),
Ok(ArtificialMessage::Permission { .. })
));
Ok(())
}
@@ -5,6 +5,7 @@ use codex_core::protocol::EventMsg;
use codex_core::protocol::Op;
use codex_core::protocol::SandboxPolicy;
use codex_execpolicy::Policy;
use codex_protocol::artificial_messages::ArtificialMessage;
use codex_protocol::models::DeveloperInstructions;
use codex_protocol::user_input::UserInput;
use codex_utils_absolute_path::AbsolutePathBuf;
@@ -34,7 +35,10 @@ fn permissions_texts(input: &[serde_json::Value]) -> Vec<String> {
.first()?
.get("text")?
.as_str()?;
if text.contains("<permissions instructions>") {
if matches!(
ArtificialMessage::parse(text),
Ok(ArtificialMessage::Permission { .. })
) {
Some(text.to_string())
} else {
None
+9 -6
View File
@@ -5,6 +5,7 @@ use anyhow::Result;
use codex_core::protocol::AskForApproval;
use codex_core::protocol::Op;
use codex_core::protocol::SandboxPolicy;
use codex_protocol::artificial_messages::ArtificialMessage;
use codex_protocol::user_input::UserInput;
use core_test_support::responses::ev_assistant_message;
use core_test_support::responses::ev_completed;
@@ -91,13 +92,15 @@ async fn user_turn_includes_skill_instructions() -> Result<()> {
let request = mock.single_request();
let user_texts = request.message_input_texts("user");
let skill_path_str = skill_path.to_string_lossy();
let skill_contents = fs::read_to_string(&skill_path)?;
let expected_skill_text = ArtificialMessage::Skill {
name: "demo".to_string(),
path: skill_path_str.to_string(),
body: skill_contents,
}
.render();
assert!(
user_texts.iter().any(|text| {
text.contains("<skill>\n<name>demo</name>")
&& text.contains("<path>")
&& text.contains(skill_body)
&& text.contains(skill_path_str.as_ref())
}),
user_texts.iter().any(|text| text == &expected_skill_text),
"expected skill instructions in user input, got {user_texts:?}"
);
+1
View File
@@ -27,6 +27,7 @@ serde_with = { workspace = true, features = ["macros", "base64"] }
strum = { workspace = true }
strum_macros = { workspace = true }
sys-locale = { workspace = true }
thiserror = { workspace = true }
tracing = { workspace = true }
ts-rs = { workspace = true, features = [
"uuid-impl",
@@ -0,0 +1,151 @@
//! Artificial message helpers shared between prompt construction and protocol parsing.
//!
//! The `artificial_messages!` macro invocation in this module defines:
//!
//! - The `ArtificialMessage` enum variants declared below (`Skill`, `ModelWarning`,
//! `Permission`) with their configured string fields.
//! - `ArtificialMessage::tag(&self) -> &'static str`: returns the top-level XML tag for
//! the current variant (for example `skill`).
//! - `ArtificialMessage::role(&self) -> &'static str`: returns the role that should be
//! used when this message is emitted as a `ResponseItem`.
//! - `ArtificialMessage::render(&self) -> String`: serializes the message to the tagged
//! XML-like envelope.
//! - `ArtificialMessage::parse(input: &str) -> Result<Self, ArtificialMessageParseError>`:
//! parses an envelope back into an enum variant.
//! - `ArtificialMessage::detect_tag(input: &str) -> Option<&'static str>`: returns the
//! known artificial-message tag if the input is a recognized envelope.
//! - `ArtificialMessage::is_artificial(input: &str) -> bool`: convenience boolean check
//! built on top of `detect_tag`.
//! - `ArtificialMessage::to_response_item(&self) -> ResponseItem`: converts the rendered
//! content into a message-shaped `ResponseItem`.
//!
//! This module also provides `impl From<ArtificialMessage> for ResponseItem` as a
//! convenience wrapper around `to_response_item`.
//!
use crate::models::ContentItem;
use crate::models::ResponseItem;
use thiserror::Error;
#[macro_use]
mod macros;
pub const TAG_SKILL: &str = "skill";
pub const TAG_MODEL_WARNING: &str = "warning";
pub const TAG_USER_SHELL_COMMAND: &str = "user_shell_cmd";
pub const TAG_PERMISSION: &str = "permissions_instructions";
pub const TAG_COLLABORATION_MODE: &str = "collaboration_mode";
pub const TAG_SPEC: &str = "personality_spec";
pub const TAG_EXEC_POLICY_AMENDMENT: &str = "exec_policy_amendment";
artificial_messages! {
Skill {
tag: TAG_SKILL,
role: "user",
fields: {
tagged(name, "name"),
tagged(path, "path"),
raw(body)
}
},
ModelWarning {
tag: TAG_MODEL_WARNING,
role: "user",
fields: {
raw(body)
}
},
Permission {
tag: TAG_PERMISSION,
role: "developer",
fields: {
raw(body)
}
}
}
#[derive(Debug, Clone, PartialEq, Eq, Error)]
pub enum ArtificialMessageParseError {
#[error("invalid artificial message envelope")]
InvalidEnvelope,
#[error("unknown artificial message tag: {0}")]
UnknownTopLevelTag(String),
#[error("missing field <{field_tag}> in message <{message_tag}>")]
MissingField {
message_tag: &'static str,
field_tag: &'static str,
},
}
impl From<ArtificialMessage> for ResponseItem {
fn from(value: ArtificialMessage) -> Self {
value.to_response_item()
}
}
#[cfg(test)]
mod tests {
use super::*;
use pretty_assertions::assert_eq;
#[test]
fn render_and_parse_skill_round_trip() {
let message = ArtificialMessage::Skill {
name: "demo".to_string(),
path: "skills/demo/SKILL.md".to_string(),
body: "body".to_string(),
};
let rendered = message.render();
assert_eq!(
rendered,
"<skill>\n<name>demo</name>\n<path>skills/demo/SKILL.md</path>\nbody\n</skill>"
);
let parsed = ArtificialMessage::parse(&rendered).expect("parse skill");
assert_eq!(parsed, message);
}
#[test]
fn render_and_parse_warning_round_trip() {
let message = ArtificialMessage::ModelWarning {
body: "be careful".to_string(),
};
let rendered = message.render();
assert_eq!(rendered, "<warning>be careful</warning>");
let parsed = ArtificialMessage::parse(&rendered).expect("parse warning");
assert_eq!(parsed, message);
}
#[test]
fn render_and_parse_permission_round_trip() {
let message = ArtificialMessage::Permission {
body: "policy text".to_string(),
};
let rendered = message.render();
assert_eq!(
rendered,
"<permissions_instructions>policy text</permissions_instructions>"
);
let parsed = ArtificialMessage::parse(&rendered).expect("parse permission");
assert_eq!(parsed, message);
}
#[test]
fn detect_tag_returns_known_tag() {
let text = "<permissions_instructions>abc</permissions_instructions>";
assert_eq!(ArtificialMessage::detect_tag(text), Some(TAG_PERMISSION));
assert!(ArtificialMessage::is_artificial(text));
}
#[test]
fn parse_rejects_unknown_tag() {
let text = "<unknown>abc</unknown>";
let err = ArtificialMessage::parse(text).expect_err("unknown tag should fail");
assert_eq!(
err,
ArtificialMessageParseError::UnknownTopLevelTag("unknown".to_string())
);
}
}
@@ -0,0 +1,206 @@
use crate::artificial_messages::ArtificialMessageParseError;
macro_rules! render_field {
($tagged:ident, $raw:ident, tagged, $value:ident, $field_tag:literal) => {{
$tagged.push_str(&$crate::artificial_messages::macros::xml(
$field_tag, $value,
));
$tagged.push('\n');
}};
($tagged:ident, $raw:ident, raw, $value:ident) => {{
$raw = Some($value.as_str());
}};
}
macro_rules! render_fields {
($($kind:ident($value:ident $(, $field_tag:literal)?)),* $(,)?) => {{
#[allow(unused_mut, unused_assignments)]
let mut tagged = String::new();
#[allow(unused_mut, unused_assignments)]
let mut raw = None;
$(
render_field!(tagged, raw, $kind, $value $(, $field_tag)?);
)*
if tagged.is_empty() {
raw.unwrap_or_default().to_string()
} else {
let mut out = String::new();
out.push('\n');
out.push_str(&tagged);
if let Some(raw) = raw {
out.push_str(raw);
out.push('\n');
}
out
}
}};
}
macro_rules! parse_field {
($remaining:ident, $consumed_tagged:ident, $msg_tag:expr, tagged, $field:ident, $field_tag:literal) => {
let tagged_input = $remaining.strip_prefix('\n').unwrap_or($remaining);
let ($field, rest) =
$crate::artificial_messages::macros::extract_prefixed_tag(tagged_input, $field_tag)
.ok_or_else(|| ArtificialMessageParseError::MissingField {
message_tag: $msg_tag,
field_tag: $field_tag,
})?;
let $field = $field.to_string();
$remaining = rest;
if !$consumed_tagged {
$consumed_tagged = true;
}
};
($remaining:ident, $consumed_tagged:ident, $msg_tag:expr, raw, $field:ident) => {
let mut value = $remaining;
if $consumed_tagged {
value = value.strip_prefix('\n').unwrap_or(value);
value = value.strip_suffix('\n').unwrap_or(value);
}
let $field = value.to_string();
$remaining = "";
};
}
macro_rules! parse_variant {
($variant:ident, $inner:expr, $msg_tag:expr, $($kind:ident($field:ident $(, $field_tag:literal)?)),* $(,)?) => {{
#[allow(unused_mut, unused_assignments)]
let mut remaining = $inner;
#[allow(unused_mut, unused_assignments)]
let mut consumed_tagged = false;
$(
parse_field!(
remaining,
consumed_tagged,
$msg_tag,
$kind,
$field
$(, $field_tag)?
);
)*
let _ = &remaining;
let _ = consumed_tagged;
Ok(ArtificialMessage::$variant { $( $field ),* })
}};
}
macro_rules! artificial_messages {
(
$(
$variant:ident {
tag: $tag:path,
role: $role:literal,
fields: { $($kind:ident($field:ident $(, $field_tag:literal)?)),* $(,)? }
}
),* $(,)?
) => {
#[derive(Debug, Clone, PartialEq, Eq)]
pub enum ArtificialMessage {
$(
$variant { $( $field: String ),* }
),*
}
impl ArtificialMessage {
pub fn tag(&self) -> &'static str {
match self {
$( Self::$variant { .. } => $tag ),*
}
}
pub fn role(&self) -> &'static str {
match self {
$( Self::$variant { .. } => $role ),*
}
}
pub fn render(&self) -> String {
match self {
$(
Self::$variant { $( $field ),* } => {
let payload = render_fields!($($kind($field $(, $field_tag)?)),*);
$crate::artificial_messages::macros::xml($tag, payload)
}
),*
}
}
pub fn parse(input: &str) -> Result<Self, ArtificialMessageParseError> {
let (tag, inner) = $crate::artificial_messages::macros::split_outer_tag(input)?;
match tag {
$(
$tag => parse_variant!($variant, inner, $tag, $($kind($field $(, $field_tag)?)),*),
)*
unknown => Err(ArtificialMessageParseError::UnknownTopLevelTag(
unknown.to_string(),
)),
}
}
pub fn detect_tag(input: &str) -> Option<&'static str> {
let (tag, _) = $crate::artificial_messages::macros::split_outer_tag(input).ok()?;
match tag {
$( $tag => Some($tag), )*
_ => None,
}
}
pub fn is_artificial(input: &str) -> bool {
Self::detect_tag(input).is_some()
}
pub fn to_response_item(&self) -> ResponseItem {
ResponseItem::Message {
role: self.role().to_string(),
id: None,
content: vec![ContentItem::InputText {
text: self.render(),
}],
end_turn: None,
phase: None,
}
}
}
};
}
pub(crate) fn split_outer_tag(input: &str) -> Result<(&str, &str), ArtificialMessageParseError> {
let text = input.trim();
if !text.starts_with('<') {
return Err(ArtificialMessageParseError::InvalidEnvelope);
}
let Some(open_end) = text.find('>') else {
return Err(ArtificialMessageParseError::InvalidEnvelope);
};
let tag = &text[1..open_end];
if tag.is_empty() || tag.starts_with('/') || tag.chars().any(char::is_whitespace) {
return Err(ArtificialMessageParseError::InvalidEnvelope);
}
let close = format!("</{tag}>");
let Some(inner) = text[open_end + 1..].strip_suffix(&close) else {
return Err(ArtificialMessageParseError::InvalidEnvelope);
};
Ok((tag, inner))
}
pub(crate) fn extract_prefixed_tag<'a>(
text: &'a str,
field_tag: &'static str,
) -> Option<(&'a str, &'a str)> {
let open = format!("<{field_tag}>");
let close = format!("</{field_tag}>");
let after_open = text.strip_prefix(&open)?;
let end = after_open.find(&close)?;
let value = &after_open[..end];
let rest = &after_open[end + close.len()..];
Some((value, rest))
}
pub(crate) fn xml(tag: &str, payload: impl std::fmt::Display) -> String {
format!("<{tag}>{payload}</{tag}>")
}
+1
View File
@@ -1,4 +1,5 @@
pub mod account;
pub mod artificial_messages;
mod thread_id;
pub use thread_id::ThreadId;
pub mod approvals;
+5 -8
View File
@@ -8,6 +8,7 @@ use serde::Serialize;
use serde::ser::Serializer;
use ts_rs::TS;
use crate::artificial_messages::ArtificialMessage;
use crate::config_types::CollaborationMode;
use crate::config_types::SandboxMode;
use crate::protocol::AskForApproval;
@@ -336,20 +337,16 @@ impl DeveloperInstructions {
request_rule_enabled: bool,
writable_roots: Option<Vec<WritableRoot>>,
) -> Self {
let start_tag = DeveloperInstructions::new("<permissions instructions>");
let end_tag = DeveloperInstructions::new("</permissions instructions>");
start_tag
.concat(DeveloperInstructions::sandbox_text(
sandbox_mode,
network_access,
))
let body = DeveloperInstructions::sandbox_text(sandbox_mode, network_access)
.concat(DeveloperInstructions::from(
approval_policy,
exec_policy,
request_rule_enabled,
))
.concat(DeveloperInstructions::from_writable_roots(writable_roots))
.concat(end_tag)
.into_text();
DeveloperInstructions::new(ArtificialMessage::Permission { body }.render())
}
fn from_writable_roots(writable_roots: Option<Vec<WritableRoot>>) -> Self {