teach-pack
Pass
Audited by Gen Agent Trust Hub on Sep 8, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No security issues were detected in the skill instructions, test records, or evaluation data. The skill implements a structured pedagogical approach focused on evidence-based learning.
- [INDIRECT_PROMPT_INJECTION]: The skill possesses a data ingestion surface as it reads user inputs and repository content to facilitate the teaching process. This is assessed as safe within its intended educational context.
- Ingestion points: User responses to probes, source code retrieved for instruction, and historical project artifacts such as PR comments (
SKILL.md). - Boundary markers: The instructions do not define specific delimiters for separating ingested content from agent instructions, though the interaction is guided by a rigid lesson structure.
- Capability inventory: The skill is authorized to write to the
.skills/teach-pack/directory and execute system commands to verify predictions against a 'runnable system' (SKILL.md). - Sanitization: No specific sanitization routines for external data are mentioned.
- Analysis: The ingestion of untrusted data is a core requirement for the 'teaching' mission. The potential for prompt injection via these channels is managed by the agent's goal-oriented interaction loop and the restricted scope of its workspace operations.
Audit Metadata