replication
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's primary function is to analyze untrusted external data (research papers, claims, or benchmarks) which could contain malicious instructions designed to hijack the agent's behavior during the replication process.
- Ingestion points: External papers, empirical claims, and benchmark data processed by the researcher agent.
- Boundary markers: Absent; the instructions do not specify any delimiters to separate untrusted data from agent instructions.
- Capability inventory: The skill generates and executes scripts in various environments (local, virtual, cloud) and writes results to the local disk.
- Sanitization: Absent; there is no mention of filtering, escaping, or validating the input data for malicious intent before it influences plan generation or script creation.
Audit Metadata