feat(pipeline): unified run_target execution behind UNIFIED_PIPELINE (#145)
CI / Check (push) Has been skipped
CI / Detect Changes (push) Successful in 3s
CI / Deploy Agent (push) Successful in 3m47s
CI / Deploy Dashboard (push) Failing after 4m28s
CI / Deploy Docs (push) Has been skipped
CI / Deploy MCP (push) Failing after 5s

This commit was merged in pull request #145.
This commit is contained in:
2026-07-12 21:31:21 +00:00
parent 17f788858d
commit c6a02f0ac3
8 changed files with 432 additions and 1 deletions
@@ -15,6 +15,7 @@ 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::sbom::SbomScanner;
use crate::pipeline::semgrep::SemgrepScanner;
@@ -419,6 +420,149 @@ impl PipelineOrchestrator {
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! { "$set": {
"status": "completed",
"current_phase": "completed",
"new_findings_count": *count as i64,
"completed_at": mongodb::bson::DateTime::now(),
} },
)
.await?;
}
Err(e) => {
tracing::error!(target_id, error = %e, "Unified scan pipeline failed");
self.db
.scan_runs()
.update_one(
doc! { "_id": &insert.inserted_id },
doc! { "$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"
);
match target.code_artifact() {
Some(code) if code.kind == ArtifactKind::GitRepo => {
let repo = repo_view_from_target(target, code);
let new_count = self.run_pipeline(&repo, scan_run_id).await?;
self.finalize_target(target, &repo, new_count).await?;
Ok(new_count)
}
Some(_) => {
tracing::warn!(
target_id = %target_id,
"Unified pipeline: source-archive scanning not yet wired; skipping"
);
Ok(0)
}
None => {
// No code to scan. Firmware/PLC/mobile static scanners land in
// #128/#129/#130; DAST for a running URL still works when a
// DastTarget row exists (migrated targets).
tracing::info!(
target_id = %target_id,
"Unified pipeline: no code artifact; attempting DAST only"
);
self.update_phase(scan_run_id, "dast_scanning").await;
self.maybe_trigger_dast(&target_id, scan_run_id).await;
Ok(0)
}
}
}
/// 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: &TrackedRepository,
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
@@ -436,6 +580,37 @@ impl PipelineOrchestrator {
}
}
/// Build a legacy `TrackedRepository` view from an onboarded target's code
/// artifact, so the unified pipeline can reuse the existing repo pipeline. The
/// inverse of the migration's `repo_to_target`. `_id` is preserved so findings
/// and DAST lookups resolve against the same key.
fn repo_view_from_target(target: &OnboardedTarget, code: &Artifact) -> TrackedRepository {
let mut repo = TrackedRepository::new(target.name.clone(), code.source_ref.clone());
repo.id = target.id;
if let Some(git) = &code.git {
repo.default_branch = git.default_branch.clone();
repo.last_scanned_commit = git.last_scanned_commit.clone();
repo.local_path = git.local_path.clone();
}
if let Some(auth) = &code.auth {
repo.auth_token = auth.secret.clone();
repo.auth_username = auth.username.clone();
}
if let Some(it) = &target.scan_config.issue_tracker {
repo.tracker_type = it.tracker_type.clone();
repo.tracker_owner = it.owner.clone();
repo.tracker_repo = it.repo.clone();
repo.tracker_token = it.token.clone();
}
repo.scan_schedule = target.scan_schedule.clone();
repo.webhook_enabled = target.webhook_enabled;
repo.webhook_secret = target.webhook_secret.clone();
repo.findings_count = target.findings_count;
repo.created_at = target.created_at;
repo.updated_at = target.updated_at;
repo
}
/// 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"
@@ -460,3 +635,50 @@ pub(super) fn extract_base_url(git_url: &str) -> Option<String> {
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 = repo_view_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);
}
}