implement
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's primary function is to implement code based on external inputs like tracker tickets and specifications, which constitutes a surface for indirect prompt injection.\n
- Ingestion points: The workflow ingests untrusted data from user requests, external issue trackers, and project specification files (
SKILL.md,sub-skills/ticket.md).\n - Boundary markers: The skill does not define specific prompt delimiters (e.g., XML tags or triple quotes) for isolating these external inputs within the agent's context, although it advises that external content should be treated as untrusted (
sub-skills/references/components/utility/conversation.md).\n - Capability inventory: The skill utilizes file system access, test execution (
tdd), and Git lifecycle management (sub-skills/spec.md).\n - Sanitization: No explicit sanitization or filtering of injected requirement text is mandated in the instructions for the agent's prompt interpolation.\n- [EXTERNAL_DOWNLOADS]: The skill's metadata and references point to external GitHub repositories for its underlying logic and component catalogs.\n
- Evidence: References to
kostja94/pagina,kostja94/bricks, andmattpocock/skillsare used as the basis for the implementation workflows and component references.
Audit Metadata