nexus-memory-agent 1.2.1

//! Derivation service - converts raw memories into explicit observations.

use std::collections::HashSet;
use std::sync::Arc;

use nexus_core::config::AgentConfig;
use nexus_core::traits::EmbeddingService;
use nexus_core::{
    cognitive_level_from_metadata, infer_perspective, perspective_from_metadata, CognitiveLevel,
    CognitiveMetadata, Memory, MemoryCategory, MemoryLaneType, PerspectiveKey, PerspectiveSource,
};
use nexus_llm::{ChatMessage, GenerateParams, LlmClient, LlmClientJson};
use nexus_storage::models::EnqueueJobParams;
use nexus_storage::repository::{
    MemoryRepository, StoreMemoryParams, StoreMemoryWithLineageParams,
};
use tracing::{debug, info, warn};

use crate::error::AgentError;
use crate::prompts::{derive_user_prompt, DERIVE_SYSTEM_PROMPT};
use crate::util::maybe_embed;

const DERIVE_MAX_TOKENS: u32 = 4096;
const REFLECT_PERSPECTIVE_JOB: &str = "reflect_perspective";
const DIGEST_SESSION_JOB: &str = "digest_session";
const DERIVE_GENERATED_BY: &str = "derive_service";
const DERIVED_FROM_ROLE: &str = "derived_from";
const RAW_ACTIVITY_LABEL: &str = "raw-activity";
const LOW_SIGNAL_LABEL: &str = "low-signal";

pub struct DeriveService {
    config: AgentConfig,
    llm: Arc<dyn LlmClient>,
    embeddings: Option<Arc<dyn EmbeddingService>>,
}

#[derive(Debug, Clone, serde::Serialize, serde::Deserialize)]
pub struct DerivedObservation {
    pub content: String,
    pub category: String,
    pub memory_lane_type: Option<String>,
    pub labels: Vec<String>,
    pub confidence: f32,
}

#[derive(Debug, Clone, serde::Serialize, serde::Deserialize)]
struct DerivedObservationEnvelope {
    observations: Vec<DerivedObservation>,
}

impl DeriveService {
    pub fn new(
        config: AgentConfig,
        llm: Arc<dyn LlmClient>,
        embeddings: Option<Arc<dyn EmbeddingService>>,
    ) -> Self {
        Self {
            config,
            llm,
            embeddings,
        }
    }

    pub async fn derive_memory(
        &self,
        memory: &Memory,
        repo: &MemoryRepository,
    ) -> Result<Vec<i64>, AgentError> {
        self.derive_memory_with_perspective(memory, None, repo)
            .await
    }

    pub async fn derive_memory_with_perspective(
        &self,
        memory: &Memory,
        queued_perspective: Option<&PerspectiveKey>,
        repo: &MemoryRepository,
    ) -> Result<Vec<i64>, AgentError> {
        if !is_derivable_source(memory) {
            return Ok(Vec::new());
        }

        let existing_ids = existing_derived_ids(repo, memory.id).await?;
        if !existing_ids.is_empty() {
            debug!(
                memory_id = memory.id,
                derived_count = existing_ids.len(),
                "Reusing existing derived observations"
            );
            return Ok(existing_ids);
        }

        let perspective = queued_perspective
            .cloned()
            .or_else(|| perspective_from_metadata(&memory.metadata))
            .unwrap_or_else(|| {
                infer_perspective(
                    PerspectiveSource::HookIngest,
                    self.config.namespace.clone(),
                    None,
                    None,
                )
            });

        let observations = match self.derive_with_llm(memory).await {
            Ok(observations) => observations,
            Err(error) => {
                warn!(memory_id = memory.id, %error, "LLM derivation failed, using fallback");
                fallback_observations(memory)
            }
        };

        let observations = normalize_observations(observations);
        if observations.is_empty() {
            debug!(memory_id = memory.id, "No explicit observations derived");
            return Ok(Vec::new());
        }

        let mut derived_ids = Vec::with_capacity(observations.len());
        for observation in observations {
            let category = MemoryCategory::parse(&observation.category).unwrap_or(memory.category);
            let memory_lane_type = observation
                .memory_lane_type
                .as_deref()
                .and_then(MemoryLaneType::parse);
            let metadata = derive_metadata(memory, &perspective, observation.confidence);
            let (embedding, embedding_model) =
                maybe_embed(self.embeddings.as_deref(), &observation.content).await;

            let derived = repo
                .store_with_lineage(StoreMemoryWithLineageParams {
                    store: StoreMemoryParams {
                        namespace_id: memory.namespace_id,
                        content: &observation.content,
                        category: &category,
                        memory_lane_type: memory_lane_type.as_ref(),
                        labels: &observation.labels,
                        metadata: &metadata,
                        embedding: embedding.as_deref(),
                        embedding_model: embedding_model.as_deref(),
                    },
                    source_memory_ids: &[memory.id],
                    evidence_role: DERIVED_FROM_ROLE,
                })
                .await
                .map_err(|error| AgentError::Storage(error.to_string()))?;
            derived_ids.push(derived.id);
        }

        enqueue_follow_up_jobs(repo, memory, &perspective, &derived_ids, &self.config).await?;

        info!(
            memory_id = memory.id,
            derived_count = derived_ids.len(),
            "Derived explicit observations from raw memory"
        );
        Ok(derived_ids)
    }

    async fn derive_with_llm(
        &self,
        memory: &Memory,
    ) -> Result<Vec<DerivedObservation>, AgentError> {
        let params = GenerateParams {
            messages: vec![
                ChatMessage::system(DERIVE_SYSTEM_PROMPT),
                ChatMessage::user(derive_user_prompt(memory)),
            ],
            max_tokens: DERIVE_MAX_TOKENS,
            temperature: 0.1,
            json_mode: true,
        };

        let envelope: DerivedObservationEnvelope = self
            .llm
            .generate_json(params)
            .await
            .map_err(|error| AgentError::Llm(error.to_string()))?;

        Ok(envelope.observations)
    }
}

fn derive_metadata(
    source: &Memory,
    perspective: &PerspectiveKey,
    confidence: f32,
) -> serde_json::Value {
    let mut cognitive = CognitiveMetadata::new(
        CognitiveLevel::Explicit,
        perspective.observer.clone(),
        perspective.subject.clone(),
        perspective.session_key.clone(),
        DERIVE_GENERATED_BY,
    );
    cognitive.source_memory_ids = vec![source.id];
    cognitive.confidence = Some(confidence.max(0.70));
    cognitive.merge_into(&sanitized_source_metadata(source))
}

fn fallback_observations(memory: &Memory) -> Vec<DerivedObservation> {
    let summary = memory
        .metadata
        .get("agent")
        .and_then(|agent| agent.get("summary"))
        .and_then(serde_json::Value::as_str)
        .map(str::trim)
        .filter(|summary| summary.len() >= 16 && !looks_like_noise(summary))
        .map(ToString::to_string);

    let content = memory
        .content
        .split_whitespace()
        .collect::<Vec<_>>()
        .join(" ");

    let candidate = summary.unwrap_or(content);
    if candidate.is_empty() || looks_like_noise(&candidate) {
        return Vec::new();
    }

    vec![DerivedObservation {
        content: candidate,
        category: memory.category.to_string(),
        memory_lane_type: memory.memory_lane_type.as_ref().map(ToString::to_string),
        labels: explicit_labels_from_source(memory),
        confidence: 0.70,
    }]
}

fn sanitized_source_metadata(source: &Memory) -> serde_json::Value {
    let mut sanitized = serde_json::Map::new();

    if let Some(agent) = source
        .metadata
        .get("agent")
        .and_then(serde_json::Value::as_object)
    {
        let mut agent_sanitized = serde_json::Map::new();
        for key in [
            "summary",
            "entities",
            "topics",
            "importance_score",
            "source",
            "generated_by",
        ] {
            if let Some(value) = agent.get(key) {
                agent_sanitized.insert(key.to_string(), value.clone());
            }
        }
        if !agent_sanitized.is_empty() {
            sanitized.insert(
                "agent".to_string(),
                serde_json::Value::Object(agent_sanitized),
            );
        }
    }

    serde_json::Value::Object(sanitized)
}

fn explicit_labels_from_source(source: &Memory) -> Vec<String> {
    let mut labels: Vec<String> = source
        .labels
        .iter()
        .filter(|label| {
            !label.eq_ignore_ascii_case(RAW_ACTIVITY_LABEL)
                && !label.eq_ignore_ascii_case(LOW_SIGNAL_LABEL)
        })
        .cloned()
        .collect();
    dedupe_labels(&mut labels);
    labels
}

fn normalize_observations(observations: Vec<DerivedObservation>) -> Vec<DerivedObservation> {
    let mut seen = HashSet::new();
    let mut normalized = Vec::new();

    for mut observation in observations {
        observation.content = observation.content.trim().to_string();
        if observation.content.is_empty() || observation.confidence < 0.70 {
            continue;
        }

        let fingerprint = observation.content.to_lowercase();
        if !seen.insert(fingerprint) {
            continue;
        }

        observation.labels.retain(|label| {
            !label.eq_ignore_ascii_case(RAW_ACTIVITY_LABEL)
                && !label.eq_ignore_ascii_case(LOW_SIGNAL_LABEL)
        });
        dedupe_labels(&mut observation.labels);
        normalized.push(observation);
    }

    normalized
}

fn dedupe_labels(labels: &mut Vec<String>) {
    let mut seen = HashSet::new();
    labels.retain(|label| seen.insert(label.to_lowercase()));
}

fn looks_like_noise(content: &str) -> bool {
    let trimmed = content.trim();
    trimmed.starts_with('{')
        || trimmed.starts_with('[')
        || trimmed.contains("\"event_name\"")
        || trimmed.contains("\"tool_name\"")
}

fn is_derivable_source(memory: &Memory) -> bool {
    if cognitive_level_from_metadata(&memory.metadata) == CognitiveLevel::Raw {
        return true;
    }

    memory.labels.iter().any(|label| {
        label.eq_ignore_ascii_case(RAW_ACTIVITY_LABEL)
            || label.eq_ignore_ascii_case(LOW_SIGNAL_LABEL)
    }) || memory
        .metadata
        .get("raw_activity")
        .and_then(serde_json::Value::as_bool)
        .unwrap_or(false)
        || memory
            .metadata
            .get("activity")
            .and_then(|activity| activity.get("low_signal"))
            .and_then(serde_json::Value::as_bool)
            .unwrap_or(false)
}

async fn existing_derived_ids(
    repo: &MemoryRepository,
    source_memory_id: i64,
) -> Result<Vec<i64>, AgentError> {
    let mut ids: Vec<i64> = repo
        .load_lineage(source_memory_id)
        .await
        .map_err(|error| AgentError::Storage(error.to_string()))?
        .into_iter()
        .filter(|entry| entry.source_memory_id == source_memory_id)
        .map(|entry| entry.derived_memory_id)
        .collect();
    ids.sort_unstable();
    ids.dedup();
    Ok(ids)
}

async fn enqueue_follow_up_jobs(
    repo: &MemoryRepository,
    source: &Memory,
    perspective: &PerspectiveKey,
    derived_ids: &[i64],
    config: &AgentConfig,
) -> Result<(), AgentError> {
    if derived_ids.is_empty() {
        return Ok(());
    }

    let perspective_json = serde_json::to_value(perspective).ok();
    let reflect_payload = serde_json::json!({
        "source_memory_id": source.id,
        "derived_memory_ids": derived_ids,
        "agent_namespace": config.namespace,
    });

    repo.enqueue_job(EnqueueJobParams {
        namespace_id: source.namespace_id,
        job_type: REFLECT_PERSPECTIVE_JOB,
        priority: 100,
        perspective: perspective_json.as_ref(),
        payload: &reflect_payload,
    })
    .await
    .map_err(|error| AgentError::Storage(error.to_string()))?;

    if let Some(session_key) = &perspective.session_key {
        let digest_payload = serde_json::json!({
            "source_memory_id": source.id,
            "derived_memory_ids": derived_ids,
            "session_key": session_key,
            "agent_namespace": config.namespace,
        });
        repo.enqueue_job(EnqueueJobParams {
            namespace_id: source.namespace_id,
            job_type: DIGEST_SESSION_JOB,
            priority: 90,
            perspective: perspective_json.as_ref(),
            payload: &digest_payload,
        })
        .await
        .map_err(|error| AgentError::Storage(error.to_string()))?;
    }

    Ok(())
}

#[cfg(test)]
mod tests {
    use super::*;

    use std::collections::VecDeque;
    use std::sync::Mutex;

    use async_trait::async_trait;
    use nexus_core::MemoryLanePriorityType;
    use nexus_llm::GenerateResponse;
    use nexus_storage::repository::{NamespaceRepository, StoreMemoryParams};
    use sqlx::sqlite::SqlitePoolOptions;

    struct MockLlmClient {
        responses: Mutex<VecDeque<nexus_llm::Result<GenerateResponse>>>,
    }

    impl MockLlmClient {
        fn new(responses: Vec<nexus_llm::Result<GenerateResponse>>) -> Self {
            Self {
                responses: Mutex::new(VecDeque::from(responses)),
            }
        }
    }

    #[async_trait]
    impl LlmClient for MockLlmClient {
        async fn generate(&self, _params: GenerateParams) -> nexus_llm::Result<GenerateResponse> {
            self.responses
                .lock()
                .expect("mock responses poisoned")
                .pop_front()
                .expect("mock response missing")
        }

        fn provider_name(&self) -> String {
            "mock".to_string()
        }

        fn model_name(&self) -> String {
            "mock-model".to_string()
        }
    }

    async fn setup_repo() -> (sqlx::SqlitePool, MemoryRepository, i64) {
        let pool = SqlitePoolOptions::new()
            .max_connections(1)
            .connect("sqlite::memory:")
            .await
            .unwrap();
        nexus_storage::migrations::run_migrations(&pool)
            .await
            .unwrap();
        let namespace_repo = NamespaceRepository::new(pool.clone());
        let namespace = namespace_repo
            .get_or_create("derive-test", "derive-test")
            .await
            .unwrap();
        let repo = MemoryRepository::new(pool.clone());
        (pool, repo, namespace.id)
    }

    fn raw_memory_metadata(session_key: &str) -> serde_json::Value {
        let mut cognitive = CognitiveMetadata::new(
            CognitiveLevel::Raw,
            "claude-code",
            "claude-code",
            Some(session_key.to_string()),
            "ingest_service",
        );
        cognitive.confidence = Some(0.9);
        cognitive.merge_into(&serde_json::json!({
            "agent": {
                "summary": "Fixed the query path and tightened pagination behavior."
            }
        }))
    }

    async fn store_raw_memory(repo: &MemoryRepository, namespace_id: i64, content: &str) -> Memory {
        repo.store(StoreMemoryParams {
            namespace_id,
            content,
            category: &MemoryCategory::Session,
            memory_lane_type: Some(&MemoryLaneType::Priority(MemoryLanePriorityType::Decision)),
            labels: &["memory".to_string(), "memory".to_string()],
            metadata: &raw_memory_metadata("session-1"),
            embedding: None,
            embedding_model: None,
        })
        .await
        .unwrap()
    }

    #[tokio::test]
    async fn test_derive_memory_skips_non_raw_memories() {
        let (_pool, repo, namespace_id) = setup_repo().await;
        let mut cognitive = CognitiveMetadata::new(
            CognitiveLevel::Explicit,
            "claude-code",
            "claude-code",
            Some("session-1".to_string()),
            "derive_service",
        );
        cognitive.confidence = Some(0.9);

        let explicit = repo
            .store(StoreMemoryParams {
                namespace_id,
                content: "Already explicit",
                category: &MemoryCategory::Facts,
                memory_lane_type: None,
                labels: &[],
                metadata: &cognitive.merge_into(&serde_json::json!({})),
                embedding: None,
                embedding_model: None,
            })
            .await
            .unwrap();

        let service = DeriveService::new(
            AgentConfig::default(),
            Arc::new(MockLlmClient::new(Vec::new())),
            None,
        );
        let derived_ids = service.derive_memory(&explicit, &repo).await.unwrap();
        assert!(derived_ids.is_empty());
    }

    #[tokio::test]
    async fn test_derive_memory_persists_explicit_observations_and_jobs() {
        let (pool, repo, namespace_id) = setup_repo().await;
        let raw = store_raw_memory(
            &repo,
            namespace_id,
            "Fixed query pagination and recall ranking.",
        )
        .await;
        let response = GenerateResponse {
            content: r#"{"observations":[{"content":"Fixed query pagination behavior.","category":"session","memory_lane_type":"decision","labels":["query","pagination"],"confidence":0.9}]}"#.to_string(),
            model: "mock-model".to_string(),
            usage: None,
        };
        let service = DeriveService::new(
            AgentConfig::default(),
            Arc::new(MockLlmClient::new(vec![Ok(response)])),
            None,
        );

        let derived_ids = service.derive_memory(&raw, &repo).await.unwrap();

        assert_eq!(derived_ids.len(), 1);
        let derived = repo.get_by_id(derived_ids[0]).await.unwrap().unwrap();
        assert_eq!(
            cognitive_level_from_metadata(&derived.metadata),
            CognitiveLevel::Explicit
        );
        assert_eq!(
            derived.labels,
            vec!["query".to_string(), "pagination".to_string()]
        );

        let lineage = repo.load_lineage(derived.id).await.unwrap();
        assert_eq!(lineage.len(), 1);
        assert_eq!(lineage[0].source_memory_id, raw.id);

        let reflect_jobs: i64 =
            sqlx::query_scalar("SELECT COUNT(*) FROM memory_jobs WHERE job_type = ?")
                .bind(REFLECT_PERSPECTIVE_JOB)
                .fetch_one(&pool)
                .await
                .unwrap();
        let digest_jobs: i64 =
            sqlx::query_scalar("SELECT COUNT(*) FROM memory_jobs WHERE job_type = ?")
                .bind(DIGEST_SESSION_JOB)
                .fetch_one(&pool)
                .await
                .unwrap();
        assert_eq!(reflect_jobs, 1);
        assert_eq!(digest_jobs, 1);
    }

    #[tokio::test]
    async fn test_derive_memory_is_idempotent() {
        let (pool, repo, namespace_id) = setup_repo().await;
        let raw = store_raw_memory(
            &repo,
            namespace_id,
            "Introduced working-set retrieval primitives.",
        )
        .await;
        let response = GenerateResponse {
            content: r#"{"observations":[{"content":"Added working-set retrieval primitives.","category":"facts","memory_lane_type":null,"labels":["retrieval"],"confidence":0.85}]}"#.to_string(),
            model: "mock-model".to_string(),
            usage: None,
        };
        let service = DeriveService::new(
            AgentConfig::default(),
            Arc::new(MockLlmClient::new(vec![Ok(response)])),
            None,
        );

        let first = service.derive_memory(&raw, &repo).await.unwrap();
        let second = service.derive_memory(&raw, &repo).await.unwrap();

        assert_eq!(first, second);

        let job_count: i64 = sqlx::query_scalar("SELECT COUNT(*) FROM memory_jobs")
            .fetch_one(&pool)
            .await
            .unwrap();
        assert_eq!(job_count, 2);
    }

    #[tokio::test]
    async fn test_derive_memory_falls_back_when_llm_response_is_invalid() {
        let (_pool, repo, namespace_id) = setup_repo().await;
        let raw = store_raw_memory(
            &repo,
            namespace_id,
            "Noisy raw content that still has a useful summary.",
        )
        .await;
        let bad_response = GenerateResponse {
            content: "not json".to_string(),
            model: "mock-model".to_string(),
            usage: None,
        };
        let service = DeriveService::new(
            AgentConfig::default(),
            Arc::new(MockLlmClient::new(vec![Ok(bad_response)])),
            None,
        );

        let derived_ids = service.derive_memory(&raw, &repo).await.unwrap();
        assert_eq!(derived_ids.len(), 1);

        let derived = repo.get_by_id(derived_ids[0]).await.unwrap().unwrap();
        assert!(derived.content.contains("tightened pagination behavior"));
    }

    #[tokio::test]
    async fn test_derive_memory_strips_raw_noise_markers_from_explicit_output() {
        let (_pool, repo, namespace_id) = setup_repo().await;
        let raw = repo
            .store(StoreMemoryParams {
                namespace_id,
                content: "Raw hook activity with durable summary.",
                category: &MemoryCategory::Session,
                memory_lane_type: None,
                labels: &[
                    RAW_ACTIVITY_LABEL.to_string(),
                    "query".to_string(),
                    "query".to_string(),
                ],
                metadata: &serde_json::json!({
                    "raw_activity": true,
                    "agent": { "summary": "Useful summary survives." },
                    "cognitive": {
                        "level": "raw",
                        "observer": "claude-code",
                        "subject": "claude-code",
                        "session_key": "session-1",
                        "generated_by": "ingest_service"
                    }
                }),
                embedding: None,
                embedding_model: None,
            })
            .await
            .unwrap();

        let response = GenerateResponse {
            content: r#"{"observations":[{"content":"Useful explicit observation.","category":"facts","memory_lane_type":null,"labels":["raw-activity","query"],"confidence":0.9}]}"#.to_string(),
            model: "mock-model".to_string(),
            usage: None,
        };
        let service = DeriveService::new(
            AgentConfig::default(),
            Arc::new(MockLlmClient::new(vec![Ok(response)])),
            None,
        );

        let derived_ids = service.derive_memory(&raw, &repo).await.unwrap();
        let derived = repo.get_by_id(derived_ids[0]).await.unwrap().unwrap();

        assert!(!derived
            .labels
            .iter()
            .any(|label| label == RAW_ACTIVITY_LABEL));
        assert!(derived.metadata.get("raw_activity").is_none());
        assert_eq!(
            derived.metadata["agent"]["summary"],
            serde_json::Value::String("Useful summary survives.".to_string())
        );
    }

    #[tokio::test]
    async fn test_derive_memory_produces_embeddings_when_service_provided() {
        let (_pool, repo, namespace_id) = setup_repo().await;
        let raw = store_raw_memory(
            &repo,
            namespace_id,
            "Implemented the query path with tight pagination.",
        )
        .await;

        let mock_embed = nexus_embeddings::MockEmbeddingService::new();
        let response = GenerateResponse {
            content: r#"{"observations":[{"content":"Implemented query path with pagination.","category":"facts","memory_lane_type":null,"labels":["query","pagination"],"confidence":0.85}]}"#.to_string(),
            model: "mock-model".to_string(),
            usage: None,
        };
        let service = DeriveService::new(
            AgentConfig::default(),
            Arc::new(MockLlmClient::new(vec![Ok(response)])),
            Some(Arc::new(mock_embed)),
        );

        let derived_ids = service.derive_memory(&raw, &repo).await.unwrap();
        let derived = repo.get_by_id(derived_ids[0]).await.unwrap().unwrap();

        assert!(
            derived.content_embedding.is_some(),
            "derived explicit observation should have an embedding when service is provided"
        );
        let embedding = derived.content_embedding.as_ref().unwrap();
        assert_eq!(embedding.len(), 384, "embedding dimension should be 384");
    }

    #[tokio::test]
    async fn test_derive_memory_stores_without_embedding_when_service_absent() {
        let (_pool, repo, namespace_id) = setup_repo().await;
        let raw = store_raw_memory(
            &repo,
            namespace_id,
            "Implemented the query path with tight pagination.",
        )
        .await;

        let response = GenerateResponse {
            content: r#"{"observations":[{"content":"Implemented query path with pagination.","category":"facts","memory_lane_type":null,"labels":["query","pagination"],"confidence":0.85}]}"#.to_string(),
            model: "mock-model".to_string(),
            usage: None,
        };
        let service = DeriveService::new(
            AgentConfig::default(),
            Arc::new(MockLlmClient::new(vec![Ok(response)])),
            None,
        );

        let derived_ids = service.derive_memory(&raw, &repo).await.unwrap();
        let derived = repo.get_by_id(derived_ids[0]).await.unwrap().unwrap();

        assert!(
            derived.content_embedding.is_none(),
            "derived observation should NOT have an embedding when no service provided"
        );
    }

    #[tokio::test]
    async fn test_derive_memory_accepts_low_signal_activity_sources() {
        let (_pool, repo, namespace_id) = setup_repo().await;
        let low_signal = repo
            .store(StoreMemoryParams {
                namespace_id,
                content: "Low signal activity with useful summary.",
                category: &MemoryCategory::Session,
                memory_lane_type: None,
                labels: &[LOW_SIGNAL_LABEL.to_string()],
                metadata: &serde_json::json!({
                    "activity": { "low_signal": true },
                    "agent": { "summary": "Captured meaningful work despite low signal." },
                    "cognitive": {
                        "level": "explicit",
                        "observer": "claude-code",
                        "subject": "claude-code",
                        "session_key": "session-1",
                        "generated_by": "activity_distiller"
                    }
                }),
                embedding: None,
                embedding_model: None,
            })
            .await
            .unwrap();

        let service = DeriveService::new(
            AgentConfig::default(),
            Arc::new(MockLlmClient::new(vec![Err(
                nexus_llm::LlmError::InvalidJsonResponse("bad".to_string()),
            )])),
            None,
        );

        let derived_ids = service.derive_memory(&low_signal, &repo).await.unwrap();
        assert_eq!(derived_ids.len(), 1);
    }
}