vectorcode
Pass
Audited by Gen Agent Trust Hub on Apr 24, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill facilitates semantic search over local codebases, which creates a surface for indirect prompt injection. Malicious instructions embedded in indexed source files could potentially influence agent behavior when retrieved during a query.
- Ingestion points: Code content is indexed via
mcp__vectorcode__vectoriseand retrieved throughmcp__vectorcode__queryin SKILL.md. - Boundary markers: No explicit instructions or delimiters are provided to the agent to treat search results as untrusted or to ignore embedded instructions in SKILL.md.
- Capability inventory: The skill uses MCP tools to interact with the file system and a local database, and documentation suggests manual usage of the
vectorcodeCLI tool in SKILL.md. - Sanitization: No sanitization or filtering of the indexed/retrieved code content is described in the skill instructions in SKILL.md.
Audit Metadata