Files
compliance-scanner-agent/compliance-agent/src/pipeline/orchestrator.rs
T
Sharang ParnerkarandClaude Fable 5 1ae6025286
CI / Check (pull_request) Failing after 1m31s
CI / Detect Changes (pull_request) Has been skipped
CI / Deploy Agent (pull_request) Has been skipped
CI / Deploy Dashboard (pull_request) Has been skipped
CI / Deploy Docs (pull_request) Has been skipped
CI / Deploy MCP (pull_request) Has been skipped
feat(plc): ephemeral soft-PLC provisioning + program load (#183)
Dynamic PLC testing without reaching the customer's device: when a PLC/SPS
target ships control logic but no reachable live URL, instantiate that logic
ourselves on a throwaway OpenPLC container in-cluster, load + start it, probe
the provisioned Modbus endpoint, and tear it down. No customer network access,
sandboxed, and reproducible.

This is the phase-1 foundation of epic #183 (OpenPLC substrate). It covers:

- provision: ephemeral container lifecycle (docker CLI). Resource-capped
  (memory/cpus/pids), hardened (no-new-privileges), labelled, joined to the
  agent's own network with no host port exposure, and swept by a stale reaper
  for anything a crashed run leaks. The `docker` argv is built by pure functions
  so it is unit-tested without a daemon.
- openplc: drives the OpenPLC web UI to load a program — login → upload →
  save → compile (MatIEC) → start_plc (which opens Modbus/TCP 502).
- runtime::provision_and_test: composes them under a hard deadline with
  guaranteed teardown on every path (success / error / timeout), then runs the
  existing ICS probe against the provisioned endpoint. extract_program picks the
  best loadable program (complete ST > largest ST > PLCopen XML).
- orchestrator: for a PlcSps target with control logic and no live URL, run
  provision-and-test after the static PLC scan. Gated by PlcRuntimeConfig
  (PLC_RUNTIME_ENABLED, default off — needs Docker access in the agent).

DAST-against-WebVisu and CODESYS-runtime fidelity are follow-ups.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-17 00:39:54 +02:00

1082 lines
41 KiB
Rust

use std::sync::Arc;
use mongodb::bson::doc;
use tracing::Instrument;
use compliance_core::models::*;
use compliance_core::traits::Scanner;
use compliance_core::AgentConfig;
use crate::database::Database;
use crate::error::AgentError;
use crate::llm::LlmClient;
use crate::pipeline::cve::CveScanner;
use crate::pipeline::git::GitOps;
use crate::pipeline::gitleaks::GitleaksScanner;
use crate::pipeline::lint::LintScanner;
use crate::pipeline::patterns::{GdprPatternScanner, OAuthPatternScanner};
use crate::pipeline::plan::build_scan_plan;
use crate::pipeline::repo_view::RepoView;
use crate::pipeline::sbom::SbomScanner;
use crate::pipeline::semgrep::SemgrepScanner;
/// Context from graph analysis passed to LLM triage for enhanced filtering
#[derive(Debug)]
#[allow(dead_code)]
pub struct GraphContext {
pub node_count: u32,
pub edge_count: u32,
pub community_count: u32,
pub impacts: Vec<compliance_core::models::graph::ImpactAnalysis>,
}
pub struct PipelineOrchestrator {
pub(super) config: AgentConfig,
pub(super) db: Database,
pub(super) llm: Arc<LlmClient>,
pub(super) http: reqwest::Client,
}
impl PipelineOrchestrator {
pub fn new(
config: AgentConfig,
db: Database,
llm: Arc<LlmClient>,
http: reqwest::Client,
) -> Self {
Self {
config,
db,
llm,
http,
}
}
#[tracing::instrument(skip_all, fields(repo_id = repo.name.as_str()))]
async fn run_pipeline(&self, repo: &RepoView, scan_run_id: &str) -> Result<u32, AgentError> {
let repo_id = repo.id.as_ref().map(|id| id.to_hex()).unwrap_or_default();
// Stage 0: Change detection
tracing::info!("[{repo_id}] Stage 0: Change detection");
let creds = GitOps::make_repo_credentials(&self.config, repo);
let git_ops = GitOps::new(&self.config.git_clone_base_path, creds);
let repo_path = git_ops.clone_or_fetch(&repo.git_url, &repo.name)?;
if !GitOps::has_new_commits(&repo_path, repo.last_scanned_commit.as_deref())? {
tracing::info!("[{repo_id}] No new commits, skipping scan");
return Ok(0);
}
let mut all_findings: Vec<Finding> = Vec::new();
// Stage 1: Semgrep SAST
tracing::info!("[{repo_id}] Stage 1: Semgrep SAST");
self.update_phase(scan_run_id, "sast").await;
match async {
let semgrep = SemgrepScanner;
semgrep.scan(&repo_path, &repo_id).await
}
.instrument(tracing::info_span!("stage_sast"))
.await
{
Ok(output) => all_findings.extend(output.findings),
Err(e) => tracing::warn!("[{repo_id}] Semgrep failed: {e}"),
}
// Stage 2: SBOM Generation
tracing::info!("[{repo_id}] Stage 2: SBOM Generation");
self.update_phase(scan_run_id, "sbom_generation").await;
let mut sbom_entries = match async {
let sbom_scanner = SbomScanner;
sbom_scanner.scan(&repo_path, &repo_id).await
}
.instrument(tracing::info_span!("stage_sbom_generation"))
.await
{
Ok(output) => output.sbom_entries,
Err(e) => {
tracing::warn!("[{repo_id}] SBOM generation failed: {e}");
Vec::new()
}
};
// Stage 3: CVE Scanning
tracing::info!("[{repo_id}] Stage 3: CVE Scanning");
self.update_phase(scan_run_id, "cve_scanning").await;
let cve_scanner = CveScanner::new(
self.http.clone(),
self.config.searxng_url.clone(),
self.config.nvd_api_key.as_ref().map(|k| {
use secrecy::ExposeSecret;
k.expose_secret().to_string()
}),
);
let cve_alerts = match tokio::time::timeout(
std::time::Duration::from_secs(600),
async {
cve_scanner
.scan_dependencies(&repo_id, &mut sbom_entries)
.await
}
.instrument(tracing::info_span!("stage_cve_scanning")),
)
.await
{
Ok(Ok(alerts)) => alerts,
Ok(Err(e)) => {
tracing::warn!("[{repo_id}] CVE scanning failed: {e}");
Vec::new()
}
Err(_) => {
tracing::warn!("[{repo_id}] CVE scanning timed out after 10 minutes");
Vec::new()
}
};
// Stage 4: Pattern Scanning (GDPR + OAuth)
tracing::info!("[{repo_id}] Stage 4: Pattern Scanning");
self.update_phase(scan_run_id, "pattern_scanning").await;
{
let pattern_findings = async {
let mut findings = Vec::new();
let gdpr = GdprPatternScanner::new();
match gdpr.scan(&repo_path, &repo_id).await {
Ok(output) => findings.extend(output.findings),
Err(e) => tracing::warn!("[{repo_id}] GDPR pattern scan failed: {e}"),
}
let oauth = OAuthPatternScanner::new();
match oauth.scan(&repo_path, &repo_id).await {
Ok(output) => findings.extend(output.findings),
Err(e) => tracing::warn!("[{repo_id}] OAuth pattern scan failed: {e}"),
}
findings
}
.instrument(tracing::info_span!("stage_pattern_scanning"))
.await;
all_findings.extend(pattern_findings);
}
// Stage 4a: Secret Detection (Gitleaks)
tracing::info!("[{repo_id}] Stage 4a: Secret Detection");
self.update_phase(scan_run_id, "secret_detection").await;
match async {
let gitleaks = GitleaksScanner;
gitleaks.scan(&repo_path, &repo_id).await
}
.instrument(tracing::info_span!("stage_secret_detection"))
.await
{
Ok(output) => all_findings.extend(output.findings),
Err(e) => tracing::warn!("[{repo_id}] Gitleaks failed: {e}"),
}
// Stage 4b: Lint Scanning
tracing::info!("[{repo_id}] Stage 4b: Lint Scanning");
self.update_phase(scan_run_id, "lint_scanning").await;
match async {
let lint = LintScanner;
lint.scan(&repo_path, &repo_id).await
}
.instrument(tracing::info_span!("stage_lint_scanning"))
.await
{
Ok(output) => all_findings.extend(output.findings),
Err(e) => tracing::warn!("[{repo_id}] Lint scanning failed: {e}"),
}
// Stage 4.5: Graph Building
tracing::info!("[{repo_id}] Stage 4.5: Graph Building");
self.update_phase(scan_run_id, "graph_building").await;
let graph_context = match async {
self.build_code_graph(&repo_path, &repo_id, &all_findings)
.await
}
.instrument(tracing::info_span!("stage_graph_building"))
.await
{
Ok(ctx) => Some(ctx),
Err(e) => {
tracing::warn!("[{repo_id}] Graph building failed: {e}");
None
}
};
// Stage 5: LLM Triage (enhanced with graph context)
tracing::info!(
"[{repo_id}] Stage 5: LLM Triage ({} findings)",
all_findings.len()
);
self.update_phase(scan_run_id, "llm_triage").await;
let triaged = crate::llm::triage::triage_findings(
&self.llm,
&mut all_findings,
graph_context.as_ref(),
)
.await;
tracing::info!("[{repo_id}] Triaged: {triaged} findings passed confidence threshold");
// Dedup against existing findings and insert new ones
let mut new_count = 0u32;
let mut new_findings: Vec<Finding> = Vec::new();
for mut finding in all_findings {
finding.scan_run_id = Some(scan_run_id.to_string());
// Check if fingerprint already exists
let existing = self
.db
.findings()
.find_one(doc! { "fingerprint": &finding.fingerprint })
.await?;
if existing.is_none() {
let result = self.db.findings().insert_one(&finding).await?;
finding.id = result.inserted_id.as_object_id();
new_findings.push(finding);
new_count += 1;
}
}
// Remove stale SBOM entries for this repo before reinserting
if !sbom_entries.is_empty() {
self.db
.sbom_entries()
.delete_many(doc! { "repo_id": &repo.id })
.await?;
}
// Persist SBOM entries
for entry in &sbom_entries {
let filter = doc! {
"repo_id": &entry.repo_id,
"name": &entry.name,
"version": &entry.version,
};
let update = mongodb::bson::to_document(entry)
.map(|d| doc! { "$set": d })
.unwrap_or_else(|_| doc! {});
self.db
.sbom_entries()
.update_one(filter, update)
.upsert(true)
.await?;
}
// Persist CVE alerts and create notifications (shared with the PLC path).
let new_notif_count = self
.persist_cve_alerts(&repo_id, &repo.name, &cve_alerts)
.await?;
if new_notif_count > 0 {
tracing::info!("[{repo_id}] Created {new_notif_count} CVE notification(s)");
}
// Stage 6: Issue Creation
tracing::info!("[{repo_id}] Stage 6: Issue Creation");
self.update_phase(scan_run_id, "issue_creation").await;
if let Err(e) = self
.create_tracker_issues(repo, &repo_id, &new_findings)
.await
{
tracing::warn!("[{repo_id}] Issue creation failed: {e}");
}
// The onboarded target's findings_count and the git artifact's
// last_scanned_commit watermark are persisted by `finalize_target` after
// `run_pipeline` returns.
// Stage 8: DAST (async, optional — only if a DastTarget is configured)
tracing::info!("[{repo_id}] Stage 8: Checking for DAST targets");
self.update_phase(scan_run_id, "dast_scanning").await;
self.maybe_trigger_dast(&repo_id, scan_run_id).await;
tracing::info!("[{repo_id}] Scan complete: {new_count} new findings");
Ok(new_count)
}
/// Unified entry point (behind `UNIFIED_PIPELINE`): run a scan for an
/// `OnboardedTarget`. Mirrors [`Self::run`] but sources the target from
/// `onboarded_targets` and dispatches by the scan plan.
#[tracing::instrument(skip_all, fields(target_id = %target_id, trigger = ?trigger))]
pub async fn run_target(
&self,
target_id: &str,
trigger: ScanTrigger,
) -> Result<(), AgentError> {
let oid = mongodb::bson::oid::ObjectId::parse_str(target_id)
.map_err(|e| AgentError::Other(e.to_string()))?;
let target = self
.db
.onboarded_targets()
.find_one(doc! { "_id": oid })
.await?
.ok_or_else(|| AgentError::Other(format!("Onboarded target {target_id} not found")))?;
let scan_run = ScanRun::new(target_id.to_string(), trigger);
let insert = self.db.scan_runs().insert_one(&scan_run).await?;
let scan_run_id = insert
.inserted_id
.as_object_id()
.map(|id| id.to_hex())
.unwrap_or_default();
let result = self.run_target_pipeline(&target, &scan_run_id).await;
match &result {
Ok(count) => {
self.db
.scan_runs()
.update_one(
doc! { "_id": &insert.inserted_id },
doc! { "$set": {
"status": "completed",
"current_phase": "completed",
"new_findings_count": *count as i64,
"completed_at": mongodb::bson::DateTime::now(),
} },
)
.await?;
// Refresh the target's cached findings count. The shared pipeline
// (Stage 7) increments `repositories`, which the unified path does
// not use, so set the accurate total on the target itself.
let total = self
.db
.findings()
.count_documents(doc! { "repo_id": target_id })
.await
.unwrap_or(*count as u64);
self.db
.onboarded_targets()
.update_one(
doc! { "_id": oid },
doc! { "$set": {
"findings_count": total as i64,
"updated_at": mongodb::bson::DateTime::now(),
} },
)
.await?;
}
Err(e) => {
tracing::error!(target_id, error = %e, "Unified scan pipeline failed");
self.db
.scan_runs()
.update_one(
doc! { "_id": &insert.inserted_id },
doc! { "$set": {
"status": "failed",
"error_message": e.to_string(),
"completed_at": mongodb::bson::DateTime::now(),
} },
)
.await?;
}
}
result.map(|_| ())
}
/// Run the applicable scans for a target. For a code target this reuses the
/// full legacy pipeline over the code artifact (clone → SAST umbrella →
/// triage → persist → issues → DAST); firmware/PLC/mobile scanners are
/// follow-ups (#128/#129/#130). Returns the number of new findings.
async fn run_target_pipeline(
&self,
target: &OnboardedTarget,
scan_run_id: &str,
) -> Result<u32, AgentError> {
let target_id = target.id.map(|id| id.to_hex()).unwrap_or_default();
let plan = build_scan_plan(target);
tracing::info!(
target_id = %target_id,
target_type = %target.target_type,
planned_steps = plan.steps.len(),
"Unified pipeline: scan plan built"
);
// Ingest + classify (tramiton for firmware) and store the detected type.
self.classify_and_store(target, &target_id, scan_run_id)
.await;
// Provision a DAST target from a LiveUrl artifact so DAST fires for
// wizard-created targets, not just migrated ones.
self.ensure_dast_target(target, &plan).await;
// PLC/SPS targets: the control-logic scan consumes the PLC source (an
// uploaded PlcProject *or* a git repo / source archive of PLCopen XML / ST
// exports), so it takes over the code artifact — we don't also run the
// SAST pipeline over it. A PLC device is reachable, so DAST still runs
// against a WebVisu / exposed endpoint when one is provisioned.
let mut new_count = 0u32;
let plc = plan.has(ScanType::PlcControlLogic);
let ics = plan.has(ScanType::IcsProbe);
if plc {
new_count += self.run_plc_scan(target, &target_id, scan_run_id).await?;
// Provision-and-test (#183): with the control logic but no reachable
// device, instantiate it on an ephemeral soft-PLC and probe that
// instead of the customer's OT network. Opt-in (needs Docker) and only
// when there is no live URL to probe directly. Never fails the scan.
if self.config.plc_runtime.enabled && target.live_url().is_none() {
match self
.run_provisioned_plc_test(target, &target_id, scan_run_id)
.await
{
Ok(n) => new_count += n,
Err(e) => {
tracing::warn!(target_id = %target_id, error = %e, "provision-and-test failed")
}
}
}
}
if ics {
new_count += self.run_ics_probe(target, &target_id, scan_run_id).await?;
}
if plc || ics {
// PLC/SPS device: also DAST against a WebVisu / exposed endpoint, but
// only when DAST is actually planned — a device reachable only over an
// industrial protocol (e.g. modbus://) has no web surface to crawl, and
// running DAST there just fails at reconnaissance. Gating here (not only
// at provisioning) also stops a DAST target left over from an earlier
// run from re-triggering. The control-logic scan already consumed the
// code artifact, so the SAST pipeline is not re-run.
if plan.has(ScanType::Dast) {
self.update_phase(scan_run_id, "dast_scanning").await;
self.maybe_trigger_dast(&target_id, scan_run_id).await;
}
return Ok(new_count);
}
match target.code_artifact() {
Some(code) if code.kind == ArtifactKind::GitRepo => {
let repo = RepoView::from_target(target, code);
let n = self.run_pipeline(&repo, scan_run_id).await?;
self.finalize_target(target, &repo, n).await?;
new_count += n;
}
Some(_) => {
tracing::warn!(
target_id = %target_id,
"Unified pipeline: source-archive scanning not yet wired; skipping"
);
}
None => {
// No code to scan (a migrated DAST target). Firmware/mobile static
// scanners land in #128/#129; DAST for a running URL works when a
// DastTarget row exists (provisioned above from a LiveUrl, or from
// a migrated target).
tracing::info!(
target_id = %target_id,
"Unified pipeline: no code artifact; attempting DAST"
);
self.update_phase(scan_run_id, "dast_scanning").await;
self.maybe_trigger_dast(&target_id, scan_run_id).await;
}
}
Ok(new_count)
}
/// Analyze a PLC/SPS project (Structured Text / PLCopen XML) for
/// control-logic security issues and persist the new findings.
async fn run_plc_scan(
&self,
target: &OnboardedTarget,
target_id: &str,
scan_run_id: &str,
) -> Result<u32, AgentError> {
tracing::info!(target_id, "[{target_id}] PLC control-logic analysis");
self.update_phase(scan_run_id, "plc_analysis").await;
let ctx = crate::ingest::IngestContext::from_config(&self.config, target_id);
let ingest_set = crate::ingest::ingest_all(target, &ctx)?;
// Every PLC-source artifact on the target: dedicated PLC projects plus any
// code artifacts (git repo / source archive) holding PLCopen XML / ST
// exports. A target can carry several (e.g. one POU export per file).
let sources: Vec<&Artifact> = target
.artifacts
.iter()
.filter(|a| {
matches!(
a.kind,
ArtifactKind::PlcProject | ArtifactKind::GitRepo | ArtifactKind::SourceArchive
)
})
.collect();
if sources.is_empty() {
tracing::warn!(target_id, "PLC scan: no PLC source artifact");
return Ok(0);
}
let mut all_findings = Vec::new();
let mut all_sbom: Vec<SbomEntry> = Vec::new();
let mut sbom_seen = std::collections::BTreeSet::new();
for a in &sources {
let Some(path) = ingest_set.get(&a.id).and_then(|ia| ia.working_path.clone()) else {
continue;
};
all_findings.extend(crate::pipeline::plc::analyze_tree(&path, target_id));
// Control-application SBOM: CODESYS libraries + runtime from a
// `.projectarchive` (uploaded, or committed in the working tree).
let archive = a
.stored_path
.clone()
.unwrap_or_else(|| a.source_ref.clone());
for e in crate::pipeline::plc::sbom::collect_sbom(
std::path::Path::new(&archive),
&path,
target_id,
) {
if sbom_seen.insert((e.name.clone(), e.version.clone())) {
all_sbom.push(e);
}
}
}
tracing::info!(
target_id,
artifacts = sources.len(),
found = all_findings.len(),
"PLC control-logic analysis complete"
);
let mut new_count = 0u32;
for mut finding in all_findings {
finding.scan_run_id = Some(scan_run_id.to_string());
if self
.db
.findings()
.find_one(doc! { "fingerprint": &finding.fingerprint })
.await?
.is_none()
{
self.db.findings().insert_one(&finding).await?;
new_count += 1;
}
}
if !all_sbom.is_empty() {
if let Err(e) = self
.persist_control_app_sbom(target_id, &target.name, all_sbom)
.await
{
tracing::warn!(target_id, error = %e, "control-app SBOM persist failed");
}
}
Ok(new_count)
}
/// Provision-and-test (#183): instantiate the target's control logic on an
/// ephemeral soft-PLC (OpenPLC), start it, probe the provisioned Modbus
/// endpoint, and tear the instance down. Used when a PLC/SPS target has the
/// control logic but no reachable live device to probe directly. Guarded by
/// `plc_runtime.enabled` (needs Docker); persists the same [`ScanType::IcsProbe`]
/// findings as a live probe.
async fn run_provisioned_plc_test(
&self,
target: &OnboardedTarget,
target_id: &str,
scan_run_id: &str,
) -> Result<u32, AgentError> {
self.update_phase(scan_run_id, "plc_provision").await;
// Locate a loadable control-logic program among the PLC-source artifacts
// (same selection as the static PLC scan: dedicated PLC projects plus code
// artifacts holding PLCopen XML / ST exports).
let ctx = crate::ingest::IngestContext::from_config(&self.config, target_id);
let ingest_set = crate::ingest::ingest_all(target, &ctx)?;
let program = target
.artifacts
.iter()
.filter(|a| {
matches!(
a.kind,
ArtifactKind::PlcProject | ArtifactKind::GitRepo | ArtifactKind::SourceArchive
)
})
.find_map(|a| {
let path = ingest_set.get(&a.id).and_then(|ia| ia.working_path.clone())?;
crate::pipeline::plc::runtime::extract_program(&path)
});
let Some(program) = program else {
tracing::info!(target_id, "provision-and-test: no loadable control-logic program");
return Ok(0);
};
let http = crate::pipeline::plc::runtime::http_client()?;
let provisioner =
crate::pipeline::plc::runtime::DockerSoftPlc::new(self.config.plc_runtime.clone());
let findings = crate::pipeline::plc::runtime::provision_and_test(
&provisioner,
&http,
&self.config.plc_runtime,
&program,
target_id,
)
.await?;
tracing::info!(
target_id,
found = findings.len(),
"provision-and-test complete"
);
let mut new_count = 0u32;
for mut finding in findings {
finding.scan_run_id = Some(scan_run_id.to_string());
if self
.db
.findings()
.find_one(doc! { "fingerprint": &finding.fingerprint })
.await?
.is_none()
{
self.db.findings().insert_one(&finding).await?;
new_count += 1;
}
}
Ok(new_count)
}
/// Probe a running PLC/SPS device over industrial protocols (Modbus/TCP, …)
/// and persist findings for exposed / unauthenticated control access. The
/// probe is read-only; it targets the Modbus port of the target's live URL.
async fn run_ics_probe(
&self,
target: &OnboardedTarget,
target_id: &str,
scan_run_id: &str,
) -> Result<u32, AgentError> {
self.update_phase(scan_run_id, "ics_probe").await;
let Some(endpoint) = target.live_url().map(|a| a.source_ref.clone()) else {
tracing::warn!(target_id, "ICS probe: no live URL");
return Ok(0);
};
// Short per-request budget so an unreachable device doesn't stall the scan.
let budget = std::time::Duration::from_secs(5);
let findings = crate::pipeline::ics::probe_target(&endpoint, target_id, budget).await;
tracing::info!(
target_id,
endpoint = %endpoint,
found = findings.len(),
"ICS probe complete"
);
let mut new_count = 0u32;
for mut finding in findings {
finding.scan_run_id = Some(scan_run_id.to_string());
if self
.db
.findings()
.find_one(doc! { "fingerprint": &finding.fingerprint })
.await?
.is_none()
{
self.db.findings().insert_one(&finding).await?;
new_count += 1;
}
}
Ok(new_count)
}
/// Store a control-application SBOM (CODESYS libraries + runtime) for a target
/// and match it against known CVEs. Scoped to `package_manager = "codesys"` so
/// it refreshes on re-scan and coexists with any firmware/source SBOM. The
/// runtime `Cmp*` / `3SLicense` components carry real CODESYS advisories, so
/// this is where PLC-device CVE coverage comes from.
async fn persist_control_app_sbom(
&self,
target_id: &str,
target_name: &str,
mut entries: Vec<SbomEntry>,
) -> Result<(), AgentError> {
if entries.is_empty() {
return Ok(());
}
self.db
.sbom_entries()
.delete_many(doc! { "repo_id": target_id, "package_manager": "codesys" })
.await?;
let cve_scanner = CveScanner::new(
self.http.clone(),
self.config.searxng_url.clone(),
self.config.nvd_api_key.as_ref().map(|k| {
use secrecy::ExposeSecret;
k.expose_secret().to_string()
}),
);
let mut alerts = match tokio::time::timeout(
std::time::Duration::from_secs(600),
cve_scanner.scan_dependencies(target_id, &mut entries),
)
.await
{
Ok(Ok(a)) => a,
Ok(Err(e)) => {
tracing::warn!(target_id, error = %e, "control-app CVE scan failed");
Vec::new()
}
Err(_) => {
tracing::warn!(target_id, "control-app CVE scan timed out");
Vec::new()
}
};
// OSV can't match `pkg:codesys/*` (no such ecosystem); CODESYS advisories
// live in NVD keyed by CPE + runtime version. Add those (best-effort).
if let Ok(codesys) = tokio::time::timeout(
std::time::Duration::from_secs(120),
cve_scanner.scan_codesys(target_id, &mut entries),
)
.await
{
alerts.extend(codesys);
} else {
tracing::warn!(target_id, "CODESYS CVE match timed out");
}
for entry in &entries {
let filter = doc! {
"repo_id": &entry.repo_id,
"name": &entry.name,
"version": &entry.version,
};
if let Ok(d) = mongodb::bson::to_document(entry) {
self.db
.sbom_entries()
.update_one(filter, doc! { "$set": d })
.upsert(true)
.await?;
}
}
let new_notifs = self
.persist_cve_alerts(target_id, target_name, &alerts)
.await?;
tracing::info!(
target_id,
components = entries.len(),
alerts = alerts.len(),
notifications = new_notifs,
"control-app SBOM stored"
);
Ok(())
}
/// Upsert CVE alerts for a target and create dedup'd CVE notifications;
/// returns the number of newly-created notifications. Shared by the SAST
/// pipeline and the PLC control-app SBOM path, so every SBOM source (source,
/// firmware, CODESYS libraries/runtime) raises the same notifications.
async fn persist_cve_alerts(
&self,
repo_id: &str,
repo_name: &str,
alerts: &[CveAlert],
) -> Result<u32, AgentError> {
use compliance_core::models::notification::{parse_severity, CveNotification};
let mut new_notif = 0u32;
for alert in alerts {
let filter = doc! { "cve_id": &alert.cve_id, "repo_id": &alert.repo_id };
let update = mongodb::bson::to_document(alert)
.map(|d| doc! { "$set": d })
.unwrap_or_else(|_| doc! {});
self.db
.cve_alerts()
.update_one(filter, update)
.upsert(true)
.await?;
// Dedup notifications by cve + repo + package + version.
let notif_filter = doc! {
"cve_id": &alert.cve_id,
"repo_id": &alert.repo_id,
"package_name": &alert.affected_package,
"package_version": &alert.affected_version,
};
let severity = parse_severity(alert.severity.as_deref(), alert.cvss_score);
let mut notification = CveNotification::new(
alert.cve_id.clone(),
repo_id.to_string(),
repo_name.to_string(),
alert.affected_package.clone(),
alert.affected_version.clone(),
severity,
);
notification.cvss_score = alert.cvss_score;
notification.summary = alert.summary.clone();
notification.url = Some(format!("https://osv.dev/vulnerability/{}", alert.cve_id));
let notif_update = doc! {
"$setOnInsert": mongodb::bson::to_bson(&notification).unwrap_or_default()
};
if let Ok(result) = self
.db
.cve_notifications()
.update_one(notif_filter, notif_update)
.upsert(true)
.await
{
if result.upserted_id.is_some() {
new_notif += 1;
}
}
}
Ok(new_notif)
}
/// Ingest the target's artifacts, classify (tramiton for firmware/RTOS/Yocto,
/// heuristics otherwise), and store the detected classification on the target.
/// Best-effort — never fails the scan.
async fn classify_and_store(
&self,
target: &OnboardedTarget,
target_id: &str,
scan_run_id: &str,
) {
self.update_phase(scan_run_id, "classification").await;
let ctx = crate::ingest::IngestContext::from_config(&self.config, target_id);
let ingest_set = match crate::ingest::ingest_all(target, &ctx) {
Ok(set) => set,
Err(e) => {
tracing::warn!(target_id, error = %e, "Unified pipeline: ingest for classification failed");
return;
}
};
let working_paths = ingest_set.working_paths();
match crate::classify::classify_target(
target,
&working_paths,
&crate::classify::TramitonNative,
)
.await
{
Ok(classification) => {
tracing::info!(
target_id,
suggested = %classification.suggested,
"Unified pipeline: classified target"
);
if let (Some(oid), Ok(bson)) = (target.id, mongodb::bson::to_bson(&classification))
{
let _ = self
.db
.onboarded_targets()
.update_one(
doc! { "_id": oid },
doc! { "$set": { "classification": bson } },
)
.await;
}
}
Err(e) => {
tracing::warn!(target_id, error = %e, "Unified pipeline: classification failed")
}
}
// Analysis-based firmware SBOM: for embedded targets, derive components
// (resolved libraries + cross-toolchain) from tramiton's build-plan
// analysis over the already-ingested source — no build, no binary
// upload. Best-effort; empty when no build plan forms.
if crate::pipeline::firmware_sbom::is_firmware_target(target.target_type) {
if let Some(code) = target.code_artifact() {
if let Some(path) = working_paths.get(&code.id) {
let entries =
crate::pipeline::firmware_sbom::firmware_sbom_entries(path, target_id)
.await;
if !entries.is_empty() {
let _ = self
.db
.sbom_entries()
.delete_many(doc! { "repo_id": target_id })
.await;
for entry in &entries {
let filter = doc! {
"repo_id": &entry.repo_id,
"name": &entry.name,
"version": &entry.version,
};
if let Ok(d) = mongodb::bson::to_document(entry) {
let _ = self
.db
.sbom_entries()
.update_one(filter, doc! { "$set": d })
.upsert(true)
.await;
}
}
tracing::info!(
target_id,
count = entries.len(),
"Firmware SBOM: stored components from tramiton analysis"
);
}
}
}
}
}
/// If the target has a `LiveUrl` artifact and DAST is planned, provision a
/// `DastTarget` (keyed by `repo_id` = target id) so the existing DAST trigger
/// fires for wizard-created targets. Idempotent.
async fn ensure_dast_target(
&self,
target: &OnboardedTarget,
plan: &crate::pipeline::plan::ScanPlan,
) {
if !plan.has(ScanType::Dast) {
return;
}
let (Some(url), Some(oid)) = (target.live_url(), target.id) else {
return;
};
let target_id = oid.to_hex();
if self
.db
.dast_targets()
.find_one(doc! { "repo_id": &target_id })
.await
.ok()
.flatten()
.is_some()
{
return; // already provisioned
}
let kind = url
.web
.as_ref()
.map(|w| w.target_kind.clone())
.unwrap_or(DastTargetType::WebApp);
let mut dast = DastTarget::new(target.name.clone(), url.source_ref.clone(), kind);
dast.repo_id = Some(target_id);
if let Some(web) = &url.web {
dast.excluded_paths = web.excluded_paths.clone();
dast.max_crawl_depth = web.max_crawl_depth;
dast.rate_limit = web.rate_limit;
dast.allow_destructive = web.allow_destructive;
}
if let Some(auth) = &url.auth {
dast.auth_config = Some(DastAuthConfig {
method: auth.method.clone(),
login_url: auth.login_url.clone(),
username: auth.username.clone(),
password: None,
token: auth.secret.clone(),
headers: auth.headers.clone(),
});
}
if let Err(e) = self.db.dast_targets().insert_one(&dast).await {
tracing::warn!(error = %e, "Unified pipeline: failed to provision DAST target");
}
}
/// Sync the onboarded-target document after a scan: bump `findings_count`
/// and advance the git artifact's `last_scanned_commit` watermark.
async fn finalize_target(
&self,
target: &OnboardedTarget,
repo: &RepoView,
new_count: u32,
) -> Result<(), AgentError> {
let oid = match target.id {
Some(id) => id,
None => return Ok(()),
};
self.db
.onboarded_targets()
.update_one(
doc! { "_id": oid },
doc! {
"$inc": { "findings_count": new_count as i64 },
"$set": { "updated_at": mongodb::bson::DateTime::now() },
},
)
.await?;
let repo_path = std::path::Path::new(&self.config.git_clone_base_path).join(&repo.name);
if let (Ok(sha), Some(code)) = (GitOps::get_head_sha(&repo_path), target.code_artifact()) {
self.db
.onboarded_targets()
.update_one(
doc! { "_id": oid, "artifacts.id": &code.id },
doc! { "$set": { "artifacts.$.git.last_scanned_commit": sha } },
)
.await?;
}
Ok(())
}
pub(super) async fn update_phase(&self, scan_run_id: &str, phase: &str) {
if let Ok(oid) = mongodb::bson::oid::ObjectId::parse_str(scan_run_id) {
let _ = self
.db
.scan_runs()
.update_one(
doc! { "_id": oid },
doc! {
"$set": { "current_phase": phase },
"$push": { "phases_completed": phase },
},
)
.await;
}
}
}
/// Extract the scheme + host from a git URL.
/// e.g. "https://gitea.example.com/owner/repo.git" -> "https://gitea.example.com"
/// e.g. "ssh://git@gitea.example.com:22/owner/repo.git" -> "https://gitea.example.com"
pub(super) fn extract_base_url(git_url: &str) -> Option<String> {
if let Some(rest) = git_url.strip_prefix("https://") {
let host = rest.split('/').next()?;
Some(format!("https://{host}"))
} else if let Some(rest) = git_url.strip_prefix("http://") {
let host = rest.split('/').next()?;
Some(format!("http://{host}"))
} else if let Some(rest) = git_url.strip_prefix("ssh://") {
// ssh://git@host:port/path -> extract host
let after_at = rest.find('@').map(|i| &rest[i + 1..]).unwrap_or(rest);
let host = after_at.split(&[':', '/'][..]).next()?;
Some(format!("https://{host}"))
} else if let Some(at_pos) = git_url.find('@') {
// SCP-style: git@host:owner/repo.git
let after_at = &git_url[at_pos + 1..];
let host = after_at.split(':').next()?;
Some(format!("https://{host}"))
} else {
None
}
}
#[cfg(test)]
#[allow(clippy::expect_used, clippy::unwrap_used)]
mod tests {
use super::*;
use compliance_core::models::{
Artifact, ArtifactAuth, IssueTrackerConfig, TargetType, TrackerType,
};
#[test]
fn repo_view_preserves_id_git_auth_and_tracker() {
let mut target = OnboardedTarget::new("acme".to_string(), TargetType::WebApp);
target.id = Some(mongodb::bson::oid::ObjectId::new());
target.findings_count = 3;
target.scan_config.issue_tracker = Some(IssueTrackerConfig {
tracker_type: Some(TrackerType::Gitea),
owner: Some("acme".to_string()),
repo: Some("web".to_string()),
token: Some("tt".to_string()),
});
let mut artifact = Artifact::git_repo("https://git/acme.git", "develop");
if let Some(git) = artifact.git.as_mut() {
git.last_scanned_commit = Some("abc123".to_string());
}
artifact.auth = Some(ArtifactAuth {
method: "token".to_string(),
username: Some("bob".to_string()),
secret: Some("pat".to_string()),
..Default::default()
});
target.artifacts.push(artifact);
let code = target.code_artifact().expect("code artifact");
let repo = RepoView::from_target(&target, code);
assert_eq!(repo.id, target.id); // preserved
assert_eq!(repo.git_url, "https://git/acme.git");
assert_eq!(repo.default_branch, "develop");
assert_eq!(repo.last_scanned_commit.as_deref(), Some("abc123"));
assert_eq!(repo.auth_token.as_deref(), Some("pat"));
assert_eq!(repo.auth_username.as_deref(), Some("bob"));
assert_eq!(repo.tracker_type, Some(TrackerType::Gitea));
assert_eq!(repo.tracker_owner.as_deref(), Some("acme"));
assert_eq!(repo.findings_count, 3);
}
}