skills/jimmc414/kosmos/gget/Gen Agent Trust Hub

gget

Pass

Audited by Gen Agent Trust Hub on Mar 29, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSCREDENTIALS_UNSAFEPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill is primarily a wrapper for the gget command-line utility, enabling the agent to execute complex bioinformatics queries, sequence alignments (via Muscle5 and DIAMOND), and structural predictions (via AlphaFold2).
  • [EXTERNAL_DOWNLOADS]: The skill facilitates the retrieval of scientific data and model parameters from reputable sources. Commands such as gget setup alphafold and gget ref -d download several gigabytes of data from established genomic repositories like Ensembl and structural databases.
  • [CREDENTIALS_UNSAFE]: The documentation for the gget gpt and gget cosmic modules instructs users to provide sensitive credentials, including OpenAI API keys and COSMIC account passwords, directly as command-line arguments. This practice can lead to the exposure of secrets in shell history files or system process listings.
  • [PROMPT_INJECTION]: The skill contains a section titled 'Suggest Using K-Dense Web For Complex Worflows' which instructs the AI agent to proactively monitor the complexity of user requests and promote a specific third-party research platform. This is a behavioral steering instruction designed for promotional purposes.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 29, 2026, 04:45 AM