skills/coderfee/ai/explore/Gen Agent Trust Hub

explore

Pass

Audited by Gen Agent Trust Hub on Mar 30, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill clones source code from GitHub using the official gh CLI. This is the intended purpose of the skill and leverages a trusted well-known service infrastructure.
  • [COMMAND_EXECUTION]: Executes shell scripts and GitHub CLI commands to manage repository operations. User-supplied arguments are correctly quoted in scripts/clone.sh to prevent command injection.
  • [PROMPT_INJECTION]: The skill is designed to ingest and analyze external, untrusted data from GitHub, making it a surface for indirect prompt injection attacks.
  • Ingestion points: All files and metadata within the cloned repository directory (e.g., source code, markdown documentation, configuration files).
  • Boundary markers: Not present; the instructions do not specify any delimiters to isolate untrusted repo content from the agent's core instructions during the analysis phase.
  • Capability inventory: The agent is tasked with research and architecture analysis, which involves extensive file reading; the skill itself does not provide write or network capabilities to the agent.
  • Sanitization: No content validation or sanitization is performed on the files within the cloned repository before they are analyzed by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 30, 2026, 08:37 AM