work-tree-canvas
Pass
Audited by Gen Agent Trust Hub on Jun 17, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local shell commands to identify the workspace and maintain data integrity. Evidence: uses
git rev-parse --show-toplevelto locate the project root,find ~to discover existing Obsidian vaults, anddateto generate precise timestamps for work nodes. - [SAFE]: The skill provides attribution and references external specifications from a well-known source. Evidence: includes a link to the
kepano/obsidian-skillsrepository for the JSON Canvas 1.0 specification. - [SAFE]: The skill processes user-controlled data files which creates an indirect prompt injection surface. 1. Ingestion points: Reads JSON data from local
.canvasfiles and a localregistry.jsonconfiguration file. 2. Boundary markers: No explicit markers are used to delimit external data. 3. Capability inventory: The skill can perform local file writes and execute specific shell commands for environment discovery. 4. Sanitization: Risk is mitigated by the highly structured nature of the JSON Canvas format and specific field-level updates performed by the skill.
Audit Metadata