github-git-recon

Pass

Audited by Gen Agent Trust Hub on Aug 2, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill facilitates the ingestion of untrusted data from external Git repositories and APIs, creating a surface for indirect prompt injection attacks. * Ingestion points: The agent is instructed to process commit logs via git log and user activity via the GitHub events API. * Boundary markers: There are no specified delimiters or instructions provided to the agent to disregard commands embedded within the external content. * Capability inventory: The skill uses git clone and suggests scanning tools that read arbitrary file content from repositories. * Sanitization: No sanitization or validation of the retrieved external content is mentioned.
  • [COMMAND_EXECUTION]: The skill provides shell commands for cloning repositories and extracting metadata. * Evidence: git clone <repo> && git log --all --format='%an <%ae>' | sort -u in SKILL.md. * Context: These commands are used to gather intelligence from Git history as part of the skill's primary function.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 2, 2026, 09:30 AM
Security Audit — agent-trust-hub — github-git-recon