api.md

run_store — Public API Reference

Data Models

RunSummary

Field	Type	Description
run_id	str	Stable run identifier.
task	str	Task text from run_started.
status	str	Derived run status.
created_at	str	First event timestamp.
updated_at	str	Last event timestamp.
path	Path	JSONL run log path.
final_answer	Optional[str]	Final answer for completed runs.

RunStore

Field	Type	Description
root	Path	Workspace root.
readonly	bool	Blocks mutating persistence operations when true.
store_dir	Path	Run storage directory .teaagent/runs.

Functions

RunStore.list_runs(*, limit: int = 20) -> list[RunSummary]

Location: teaagent/run_store.py:125 Pre-condition: None. Post-condition: Returns newest-first run summaries; corrupt files are skipped and counted.

RunStore.show_run(run_id: str) -> list[dict[str, Any]]

Location: teaagent/run_store.py:138 Pre-condition: Run file exists. Post-condition: Returns parsed event list; malformed lines increment corruption count.

RunStore.task_for_run(run_id: str) -> str

Location: teaagent/run_store.py:152 Pre-condition: Run contains run_started with a string task. Post-condition: Returns task string or raises ValueError.

RunStore.observations_for_run(run_id: str) -> list[dict[str, Any]]

Location: teaagent/run_store.py:160 Pre-condition: None. Post-condition: Returns normalized tool_call_completed observation records.

RunStore.pending_approval_for_run(run_id: str) -> Optional[dict[str, Any]]

Location: teaagent/run_store.py:179 Pre-condition: None. Post-condition: Returns unresolved approval payload if one is pending, else None.

RunStore.health_report() -> dict[str, Any]

Location: teaagent/run_store.py:296 Pre-condition: None. Post-condition: Returns corruption-aware health snapshot.