ai-data-privacy
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted external data (code, logs, and configuration files) which may contain malicious instructions meant to hijack the agent's behavior.
- Ingestion points: Processes external files using the
Read,Grep, andGlobtools based on target files or directories provided in the skill's arguments. - Boundary markers: Includes a dedicated "Prompt Injection Safety Notice" that explicitly instructs the agent to treat reviewed content as data for analysis rather than instructions to follow, and to flag payloads as findings.
- Capability inventory: The skill is constrained by its frontmatter to a read-only environment. It lacks tools for network access, file modification, or shell command execution, significantly limiting the potential impact of any successful injection.
- Sanitization: Relies on instructional guardrails and strict tool scoping to ensure that data processed from the filesystem cannot be executed as code or interpreted as system-level commands.
Audit Metadata