reminder-review-session
Pass
Audited by Gen Agent Trust Hub on May 5, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: Executes a local script
scripts/obsidian-helper.pyusing theuv runcommand to manage Obsidian vault interactions. - [PROMPT_INJECTION]: Indirect Prompt Injection Surface:
- Ingestion points: The skill analyzes the full session conversation history to identify content worth summarizing and saving to Obsidian.
- Boundary markers: Absent. The instructions do not define delimiters or specific safety warnings to the agent to ignore instructions embedded within the processed chat history.
- Capability inventory: The skill possesses capabilities for command execution via subprocesses (
uv run) and local file system access (read/write operations on markdown files). - Sanitization: Absent. There is no evidence of validation or escaping mechanisms for the session data before it is interpolated into files or processed by the summarizing logic.
Audit Metadata