deep-work
Pass
Audited by Gen Agent Trust Hub on Nov 2, 2026
Risk Level: SAFEDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill implements deterministic logic for productivity planning using Python scripts that rely solely on the standard library.- [DATA_EXPOSURE]: The skill stores user focus session history in
~/.deep-work/sessions.json. This is local storage for application state and does not target sensitive system credentials or configurations.- [INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided task descriptions which are processed by the auditor and planner scripts. - Ingestion points: Task names and durations are accepted via CLI arguments in
shallow_work_auditor.pyandtime_block_planner.py. - Boundary markers: None.
- Capability inventory: File-write operations to
~/.deep-work/sessions.jsoninfocus_session_logger.py. - Sanitization: Task names are treated as strings for keyword matching and logging; no dynamic execution of input data is performed.
Audit Metadata