quickstart

Pass

Audited by Gen Agent Trust Hub on Oct 24, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill fetches a bootstrap shell script and a markdown runbook from basic-anteater-667.convex.site. These resources are used to initialize the project environment and provide the agent with current build rules.
  • [COMMAND_EXECUTION]: The skill executes the downloaded bootstrap script using bash /tmp/convex-qb.sh. This script is responsible for the project scaffolding and local environment setup.
  • [INDIRECT_PROMPT_INJECTION]: The agent is instructed to fetch and follow instructions from an external markdown file (runbook). This introduces a surface where instructions hosted on the vendor's server guide the agent's actions during the build phase.
  • Ingestion points: SKILL.md instructs the agent to fetch the runbook URL using its browser/web tools.
  • Boundary markers: No specific delimiters are used to separate the external runbook content from the agent's core instructions.
  • Capability inventory: The agent has capabilities to execute shell commands, write files, and invoke subagents (convex-expert).
  • Sanitization: The fetched markdown content is not sanitized before being adopted as the 'canonical' rule set.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 24, 2026, 07:59 AM