monthly-review
Pass
Audited by Gen Agent Trust Hub on Mar 14, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local CLI commands
obsidian searchandobsidian evalto query vault data and perform date calculations. - [PROMPT_INJECTION]: The skill ingests untrusted text from daily and weekly notes for thematic analysis and inferences, creating an attack surface for indirect prompt injection.
- Ingestion points: Reads content from
periodic/daily/andperiodic/weekly/files to identify themes and language patterns. - Boundary markers: Uses structural markers like callout headers (
[!weekly-review]-) to delineate data, though arbitrary text following these markers is ingested. - Capability inventory: Permissions to read multiple directories, write to monthly notes, and execute local search/eval commands.
- Sanitization: No explicit sanitization or instruction filtering is applied to the ingested text before it is used to generate inferences.
Audit Metadata