implement
Pass
Audited by Gen Agent Trust Hub on Aug 21, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest external data in the form of specifications and tickets.
- Ingestion points: User-provided specifications or tickets (SKILL.md).
- Boundary markers: None present to delineate external content.
- Capability inventory: The skill utilizes custom commands ($ponytail, $tdd, $implementation-review), interacts with browsers and emulators, runs test suites, and performs Git operations.
- Sanitization: No explicit sanitization or filtering of input data is mentioned.
- Note: This represents a standard surface for a development-focused agent and aligns with its primary purpose.
Audit Metadata