start-learning
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill fetches the
ROADMAP.mdfile from the author's GitHub repository (github.com/fancyboi999) to determine estimated hours for the curriculum phases. This is a legitimate use of external resources for curriculum data.\n- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user input during the interview process (Mission and Build Goal) and writes it directly into theLEARNING.mdfile, which serves as the persistent state for subsequent learning sessions. This creates a surface where malicious user input could potentially influence the agent's behavior during later lessons.\n - Ingestion points: Free-text responses to interview questions 1 and 3 in
SKILL.md.\n - Boundary markers: The skill does not use delimiters or instructions to ignore embedded content when writing user input to
LEARNING.md.\n - Capability inventory: The skill utilizes file system write access to create the learning plan.\n
- Sanitization: User input is recorded exactly as provided without validation or escaping.
Audit Metadata