metodyki-nauczania
Pass
Audited by Gen Agent Trust Hub on Nov 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill possesses a vulnerability surface for indirect prompt injection as it is designed to analyze untrusted user-provided didactic briefs and project files.
- Ingestion points: The agent is instructed to process user-supplied "briefs" and project context as seen in SKILL.md ("Prowadzenie twórcy przez decyzje projektowe") and TEACHERS.md.
- Boundary markers: While the skill prompt does not explicitly implement delimiters for its own input processing, it provides extensive guidance in references/runtime-ai.md on implementing these for the resulting applications.
- Capability inventory: The skill does not request specific tools in its frontmatter, but if the agent environment provides file system or network access, these could be influenced by malicious content in project files being analyzed.
- Sanitization: No sanitization is performed by the skill's instructions on its own inputs, although it provides detailed sanitization and validation advice for the applications it helps build.
- Conclusion: This risk is intrinsic to the skill's primary purpose of project analysis and is mitigated by the skill's own defensive didactic recommendations. Per internal guidelines, the severity is adjusted to SAFE.
- [REMOTE_CODE_EXECUTION]: The README.md and USAGE.md provide standard installation instructions using the 'skills CLI' pointing to the author's official GitHub repository (github.com/siberius/learning-methodologies). This is an external reference to the vendor's own infrastructure and is documented neutrally as a standard installation mechanism for the platform.
Audit Metadata