knowledge-base-query
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructions include a step to clone the knowledge base repository from
git@github.com:mmontes11/knowledge-base.gitif the local path~/code/knowledge-baseis missing. This repository is owned by the skill's author and provides the necessary documentation files for the skill's operation. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data by reading various markdown files from the cloned repository, which presents a surface for indirect prompt injection if the repository content were to be compromised.
- Ingestion points: Multiple markdown files including
README.md,api-reference.md,releases.md,features.md, and others located within theprojects/,books/, andfoundation/directories of the knowledge base. - Boundary markers: The instructions do not define specific delimiters or "ignore instructions" wrappers for the content retrieved from the files.
- Capability inventory: The skill is strictly limited to read-only capabilities, utilizing only the
Read,Glob, andGreptools. - Sanitization: No sanitization or validation of the file content is performed prior to the agent processing and summarizing the data.
Audit Metadata