Skip to content

feat: project ContextProvenance explicitly #34

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
ForeverAngry merged 1 commit into from
Jun 1, 2026
Merged

feat: project ContextProvenance explicitly #34

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
138 changes: 86 additions & 52 deletions src/projection.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,7 +1,8 @@
//! Projection helpers for `rig-compose` context packing.

use rig_compose::{
ContextItem, ContextPack, ContextPackConfig, ContextSourceKind, Evidence, InvestigationContext,
ContextItem, ContextPack, ContextPackConfig, ContextProjectionState, ContextProvenance,
ContextSourceKind, Evidence, InvestigationContext,
};
use serde_json::{Value, json};

Expand All @@ -10,10 +11,6 @@ use crate::{BehaviorPattern, EntityBaseline, MemoryLookupHit};
#[cfg(feature = "graph")]
use crate::Subgraph;

const STATE_CANDIDATE: &str = "candidate";
#[cfg(feature = "graph")]
const STATE_EXPANDED: &str = "expanded";

/// Convert resource-native records into [`ContextItem`] values.
pub trait IntoContextItem {
/// Project this resource record into a prompt-ready context item.
Expand All @@ -30,12 +27,15 @@ impl IntoContextItem for BehaviorPattern {
};
ContextItem::new(ContextSourceKind::Resource, source_id, text)
.with_score(f64::from(self.confidence_delta))
.with_provenance(json!({
.with_context_provenance(
ContextProvenance::new()
.with_source_uri(format!("behavior-pattern://{}@v{}", self.id, self.version))
.with_confidence(f64::from(self.confidence_delta))
.with_version_key(&self.id)
.with_projection_state(ContextProjectionState::Candidate),
)
.with_metadata(json!({
"resource": "behavior_pattern",
"source_uri": format!("behavior-pattern://{}@v{}", self.id, self.version),
"confidence": self.confidence_delta,
"version_key": self.id,
"projection_state": STATE_CANDIDATE,
"id": self.id,
"version": self.version,
"required": self.rule.required,
Expand All @@ -57,12 +57,15 @@ impl IntoContextItem for EntityBaseline {
),
)
.with_score(self.samples as f64)
.with_provenance(json!({
.with_context_provenance(
ContextProvenance::new()
.with_source_uri(format!("baseline://{}/{}", self.entity, self.metric))
.with_principal(&self.entity)
.with_confidence(self.samples as f64)
.with_projection_state(ContextProjectionState::Candidate),
)
.with_metadata(json!({
"resource": "baseline",
"source_uri": format!("baseline://{}/{}", self.entity, self.metric),
"principal": self.entity,
"confidence": self.samples,
"projection_state": STATE_CANDIDATE,
"entity": self.entity,
"metric": self.metric,
"mean": self.mean,
Expand Down Expand Up @@ -92,19 +95,32 @@ pub fn memory_hit_to_context_item(hit: &MemoryLookupHit, rank: usize) -> Context
.key
.clone()
.unwrap_or_else(|| format!("memory.hit/{rank}"));
let mut prov = ContextProvenance::new()
.with_projection_state(ContextProjectionState::Candidate)
.with_confidence(f64::from(hit.score));
if let Some(uri) = &hit.source_uri {
prov = prov.with_source_uri(uri);
}
if let Some(principal) = &hit.principal {
prov = prov.with_principal(principal);
}
if let Some(scope) = &hit.scope {
prov = prov.with_scope(scope);
}
if let Some(ms) = hit.recorded_at_millis {
prov = prov.with_recorded_at_millis(ms);
}
if let Some(key) = &hit.key {
prov = prov.with_source_frame_id(key);
}

ContextItem::new(ContextSourceKind::Memory, source_id, hit.summary.clone())
.with_rank(rank)
.with_score(f64::from(hit.score))
.with_provenance(json!({
.with_context_provenance(prov)
.with_metadata(json!({
"resource": "memory.lookup",
"key": hit.key,
"source_uri": hit.source_uri,
"principal": hit.principal,
"scope": hit.scope,
"recorded_at_millis": hit.recorded_at_millis,
"confidence": hit.score,
"source_frame_id": hit.key,
"projection_state": STATE_CANDIDATE,
"score": hit.score,
"metadata": hit.metadata,
}))
Expand Down Expand Up @@ -146,12 +162,15 @@ pub fn subgraph_to_context_item(subgraph: &Subgraph, rank: usize) -> ContextItem
)
.with_rank(rank)
.with_score(node_count.saturating_add(edge_count) as f64)
.with_provenance(json!({
.with_context_provenance(
ContextProvenance::new()
.with_source_uri(format!("graph://{}", subgraph.seed))
.with_principal(&subgraph.seed)
.with_projection_state(ContextProjectionState::Expanded)
.with_reason("graph_expansion"),
)
.with_metadata(json!({
"resource": "graph.subgraph",
"source_uri": format!("graph://{}", subgraph.seed),
"principal": subgraph.seed,
"projection_state": STATE_EXPANDED,
"reason": "graph_expansion",
"seed": subgraph.seed,
"nodes": subgraph.nodes,
"edges": subgraph.edges,
Expand All @@ -172,11 +191,17 @@ pub fn evidence_to_context_item(evidence: &Evidence, rank: usize) -> ContextItem
ContextItem::new(source, source_id, evidence_text(evidence))
.with_rank(rank)
.with_score(evidence_score(&evidence.detail))
.with_provenance(json!({
.with_context_provenance(
ContextProvenance::new()
.with_source_uri(format!(
"evidence://{}/{}",
evidence.source_skill, evidence.label
))
.with_confidence(evidence_score(&evidence.detail))
.with_projection_state(ContextProjectionState::Candidate),
)
.with_metadata(json!({
"resource": "investigation.evidence",
"source_uri": format!("evidence://{}/{}", evidence.source_skill, evidence.label),
"confidence": evidence_score(&evidence.detail),
"projection_state": STATE_CANDIDATE,
"source_skill": evidence.source_skill,
"label": evidence.label,
"detail": evidence.detail,
Expand Down Expand Up @@ -233,10 +258,14 @@ mod tests {
assert_eq!(item.source_id, "behavior_pattern/spray@v2");
assert_eq!(item.text, "password spray around one host");
assert!((item.score - 0.25).abs() < 1e-9);
assert_eq!(item.provenance["resource"], "behavior_pattern");
assert_eq!(item.provenance["source_uri"], "behavior-pattern://spray@v2");
assert_eq!(item.provenance["projection_state"], "candidate");
assert_eq!(item.provenance["required"][0], "auth.failure.burst");
assert_eq!(item.metadata["resource"], "behavior_pattern");
let prov = item.context_provenance().unwrap();
assert_eq!(prov.source_uri.unwrap(), "behavior-pattern://spray@v2");
assert_eq!(
prov.projection_state.unwrap(),
ContextProjectionState::Candidate
);
assert_eq!(item.metadata["required"][0], "auth.failure.burst");
}

#[test]
Expand All @@ -257,19 +286,20 @@ mod tests {
assert_eq!(items[0].source, ContextSourceKind::Memory);
assert_eq!(items[0].source_id, "episode-1");
assert_eq!(items[0].rank, 0);
assert_eq!(items[0].provenance["source_uri"], "memory://episode/1");
assert_eq!(items[0].provenance["principal"], "alice");
assert_eq!(items[0].provenance["scope"], "workspace");
assert_eq!(
items[0].provenance["recorded_at_millis"],
1_700_000_000_000_i64
);
let confidence = items[0].provenance["confidence"]
.as_f64()
let prov = items[0].context_provenance().unwrap();
assert_eq!(prov.source_uri.unwrap(), "memory://episode/1");
assert_eq!(prov.principal.unwrap(), "alice");
assert_eq!(prov.scope.unwrap(), "workspace");
assert_eq!(prov.recorded_at_millis.unwrap(), 1_700_000_000_000_i64);
let confidence = prov
.confidence
.expect("confidence should serialize as a number");
assert!((confidence - 0.9).abs() < 1e-6);
assert_eq!(items[0].provenance["source_frame_id"], "episode-1");
assert_eq!(items[0].provenance["projection_state"], "candidate");
assert_eq!(prov.source_frame_id.unwrap(), "episode-1");
assert_eq!(
prov.projection_state.unwrap(),
ContextProjectionState::Candidate
);
assert_eq!(items[1].source_id, "memory.hit/1");
assert_eq!(items[1].rank, 1);
}
Expand Down Expand Up @@ -313,10 +343,14 @@ mod tests {
assert_eq!(item.source_id, "graph/host-1");
assert_eq!(item.rank, 3);
assert_eq!(item.score, 3.0);
assert_eq!(item.provenance["resource"], "graph.subgraph");
assert_eq!(item.provenance["source_uri"], "graph://host-1");
assert_eq!(item.provenance["projection_state"], "expanded");
assert_eq!(item.provenance["reason"], "graph_expansion");
assert_eq!(item.provenance["seed"], "host-1");
assert_eq!(item.metadata["resource"], "graph.subgraph");
let prov = item.context_provenance().unwrap();
assert_eq!(prov.source_uri.unwrap(), "graph://host-1");
assert_eq!(
prov.projection_state.unwrap(),
ContextProjectionState::Expanded
);
assert_eq!(prov.reason.unwrap(), "graph_expansion");
assert_eq!(item.metadata["seed"], "host-1");
}
}
76 changes: 44 additions & 32 deletions src/security/finding.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -26,18 +26,17 @@
//! .with_technique_id("T1110.003");
//!
//! let item = security_finding_to_context_item(&finding, 0);
//! assert_eq!(item.provenance["finding_id"], "credential.password_spray");
//! assert_eq!(item.provenance["severity"], "high");
//! assert_eq!(item.metadata["finding_id"], "credential.password_spray");
//! assert_eq!(item.metadata["severity"], "high");
//! ```

use rig_compose::{ContextItem, ContextSourceKind};
use rig_compose::{ContextItem, ContextProjectionState, ContextProvenance, ContextSourceKind};
use serde::{Deserialize, Serialize};
use serde_json::{Value, json};

use crate::projection::IntoContextItem;
use crate::trace::ResourceTraceEnvelope;

const STATE_CANDIDATE: &str = "candidate";
const TRACE_RESOURCE: &str = "security";
const TRACE_OPERATION: &str = "finding";
const TRACE_KIND: &str = "security_finding";
Expand Down Expand Up @@ -275,21 +274,31 @@ pub fn security_finding_to_context_item(finding: &SecurityFinding, rank: usize)
let source_uri = finding.resolved_source_uri();
let confidence = finding.effective_confidence();

let mut prov = ContextProvenance::new()
.with_projection_state(ContextProjectionState::Candidate)
.with_confidence(confidence);

prov = prov.with_source_uri(source_uri);
if let Some(principal) = &finding.principal {
prov = prov.with_principal(principal);
}
if let Some(scope) = &finding.scope {
prov = prov.with_scope(scope);
}
if let Some(ms) = finding.recorded_at_millis {
prov = prov.with_recorded_at_millis(ms);
}

ContextItem::new(
ContextSourceKind::Resource,
source_id,
finding.summary.clone(),
)
.with_rank(rank)
.with_score(confidence)
.with_provenance(json!({
.with_context_provenance(prov)
.with_metadata(json!({
"resource": "security.finding",
"source_uri": source_uri,
"principal": finding.principal,
"scope": finding.scope,
"recorded_at_millis": finding.recorded_at_millis,
"confidence": confidence,
"projection_state": STATE_CANDIDATE,
"finding_id": finding.id,
"severity": finding.severity.as_str(),
"technique_id": finding.technique_id,
Expand Down Expand Up @@ -402,24 +411,27 @@ mod tests {
assert_eq!(item.rank, 3);
assert!((item.score - FindingSeverity::High.confidence_weight()).abs() < 1e-9);

let p = &item.provenance;
// Shared vocabulary
assert_eq!(p["resource"], "security.finding");
assert_eq!(p["source_uri"], "siem://event/42");
assert_eq!(p["principal"], "host-1");
assert_eq!(p["scope"], "workspace");
assert_eq!(p["recorded_at_millis"], 1_700_000_000_000_i64);
assert_eq!(p["projection_state"], "candidate");
let confidence = p["confidence"].as_f64().unwrap();
let m = &item.metadata;
assert_eq!(m["resource"], "security.finding");
assert_eq!(m["finding_id"], "credential.password_spray");
assert_eq!(m["severity"], "high");
assert_eq!(m["technique_id"], "T1110.003");
assert_eq!(m["tactic"], "credential-access");
assert_eq!(m["source_skill"], "credential.password_spray");
assert_eq!(m["signals"][0], "auth.failure.burst");
assert_eq!(m["detail"]["distinct_accounts"], 17);

let prov = item.context_provenance().unwrap();
assert_eq!(prov.source_uri.unwrap(), "siem://event/42");
assert_eq!(prov.principal.unwrap(), "host-1");
assert_eq!(prov.scope.unwrap(), "workspace");
assert_eq!(prov.recorded_at_millis.unwrap(), 1_700_000_000_000_i64);
assert_eq!(
prov.projection_state.unwrap(),
ContextProjectionState::Candidate
);
let confidence = prov.confidence.unwrap();
assert!((confidence - FindingSeverity::High.confidence_weight()).abs() < 1e-9);
// Security-specific
assert_eq!(p["finding_id"], "credential.password_spray");
assert_eq!(p["severity"], "high");
assert_eq!(p["technique_id"], "T1110.003");
assert_eq!(p["tactic"], "credential-access");
assert_eq!(p["source_skill"], "credential.password_spray");
assert_eq!(p["signals"][0], "auth.failure.burst");
assert_eq!(p["detail"]["distinct_accounts"], 17);
}

#[test]
Expand All @@ -429,18 +441,18 @@ mod tests {

assert_eq!(item.source_id, "security.finding/recon.high_fanout");
assert_eq!(
item.provenance["source_uri"],
item.context_provenance().unwrap().source_uri.unwrap(),
"security-finding://recon.high_fanout"
);
assert!(item.provenance["principal"].is_null());
assert!(item.context_provenance().unwrap().principal.is_none());
}

#[test]
fn explicit_confidence_overrides_severity_weight() {
let finding = SecurityFinding::new("exfil.slow_beacon", FindingSeverity::Low, "beacon")
.with_confidence(0.99);
let item = security_finding_to_context_item(&finding, 0);
let confidence = item.provenance["confidence"].as_f64().unwrap();
let confidence = item.context_provenance().unwrap().confidence.unwrap();
assert!((confidence - 0.99).abs() < 1e-9);
assert!((item.score - 0.99).abs() < 1e-9);
}
Expand All @@ -467,7 +479,7 @@ mod tests {
"chain",
);
let item: ContextItem = finding.to_context_item();
assert_eq!(item.provenance["severity"], "critical");
assert_eq!(item.metadata["severity"], "critical");
}

#[test]
Expand Down
Loading
Loading