decision-logger
Pass
Audited by Gen Agent Trust Hub on Jun 16, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [SAFE]: The skill implements a secure-by-design two-layer memory architecture for handling sensitive meeting data.
- Ingestion points: The skill processes meeting transcripts from the
memory/board-meetings/directory. - Boundary markers: Uses specific structured templates (e.g.,
templates/decision-entry.md) to define decision fields. - Capability inventory: The skill performs local file read/write operations and executes a local script
scripts/decision_tracker.py. - Sanitization: Security is maintained through a mandatory human-in-the-loop 'Founder Approval' gate before any data is promoted from the raw transcript (Layer 1) to the approved decision memory (Layer 2).
- [COMMAND_EXECUTION]: The skill defines several CLI commands to interface with its internal tracking script,
scripts/decision_tracker.py. These commands are used for administrative tasks such as searching history, detecting conflicts, and reviewing overdue action items, operating solely on local data.
Audit Metadata