knowledge-stack
Pass
Audited by Gen Agent Trust Hub on Jul 17, 2026
Risk Level: SAFEDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill instructs the agent to access and reason over sensitive personal and professional data located in local directories including
~/Obsidian/and~/Developer/. This access includes personal journals, release engineering notes, and sensitive legal or financial information such as tax amendments and housing situations. - [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface because it ingests data from external sources (GitHub Issues) and local Markdown files that could contain embedded instructions intended to override agent behavior.
- Ingestion points: Local Markdown files in multiple Obsidian vaults and remote GitHub issues via the
gh issuetool. - Boundary markers: The instructions do not specify any delimiters or boundary markers to help the agent distinguish between data and instructions within the retrieved files.
- Capability inventory: The agent is granted capabilities to search, read, and write files across the filesystem and interact with GitHub repository data.
- Sanitization: There is no mention of sanitization, filtering, or validation for the content retrieved from these various sources.
Audit Metadata