pub struct WorkflowSummary {
pub workflow_type: String,
pub steps_detected: Vec<String>,
pub missing_recommended_steps: Vec<String>,
pub evidence_quality: EvidenceQuality,
pub task_id: String,
pub plan_id: String,
pub closed_at: String,
pub symbols_annotated: usize,
pub ledger_entries_written: usize,
pub db_state: String,
pub db_state_note: String,
}Expand description
Rolled-up workflow summary emitted in task-close JSON and persisted to
.asd/workflow-sessions.jsonl.
Fields§
§workflow_type: StringOne of: “full” | “annotate-and-close” | “test-and-close” | “close-only”.
steps_detected: Vec<String>Steps that were detected from ledger evidence.
missing_recommended_steps: Vec<String>Recommended steps that were not detected.
evidence_quality: EvidenceQualityPer-field evidence quality breakdown.
task_id: StringCTX task id (empty string when absent).
plan_id: StringCTX plan id (empty string when absent).
closed_at: StringRFC-3339 timestamp of this task-close invocation.
symbols_annotated: usizeNumber of symbols that received ledger entries.
ledger_entries_written: usizeTotal ledger entries written in this task-close call.
db_state: StringData-quality state of the workspace at task-close time.
Mirrors DataQuality::state from the trust score:
"clean_room" | "unannotated" | "sparse_but_active" | "populated" | "degraded" | "empty".
Helps agents understand why evidence_quality_score is low on a fresh
workspace: a score of 0.17 on an unannotated DB is expected, not a bug.
db_state_note: StringHuman-readable clarification when db_state may need explanation.
Set for clean_room and unannotated (low evidence expected),
and for degraded (possible state loss). Empty for healthy states.