skills/coco-research/coco/find-skills/Gen Agent Trust Hub

find-skills

Pass

Audited by Gen Agent Trust Hub on Nov 7, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the npx skills CLI to interactively search for and install modular packages. This involves executing system-level commands to manage the agent's environment.
  • [EXTERNAL_DOWNLOADS]: The skill facilitates the discovery and installation of external code from GitHub and the official skills registry. It provides guidance on verifying source URLs and repository owners before proceeding with downloads.
  • [INDIRECT_PROMPT_INJECTION]: As the skill ingests search results from external sources, there is a theoretical risk of malicious metadata influencing agent behavior. The skill proactively addresses this by requiring the agent to display exact package identifiers, repository URLs, and README excerpts to the user for explicit confirmation before execution. These boundary markers and human review steps mitigate the risk of automated exploitation.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 7, 2026, 06:11 PM