configuration-env

Pass

Audited by Gen Agent Trust Hub on Nov 8, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
  • [COMMAND_EXECUTION]: The utility script scripts/check-config.sh accepts a --start argument which is executed using bash -lc. This grants the agent the capability to run arbitrary shell commands on the host environment to validate service startup.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze potentially untrusted data from configuration files such as .env, .env.example, and wrangler.toml. These files could contain hidden instructions that influence the agent's behavior.
  • Ingestion points: The agent is instructed in SKILL.md to read and compare various configuration files including .env and wrangler.* formats.
  • Boundary markers: No explicit delimiters or instructions to ignore embedded content are provided when the agent processes these configuration files.
  • Capability inventory: The skill provides access to shell execution (bash -lc) and network requests (curl) through the scripts/check-config.sh script.
  • Sanitization: There is no evidence of sanitization or content validation performed on the ingested files before they are used to generate commands or inform agent actions.
  • [DATA_EXFILTRATION]: The skill explicitly targets sensitive files such as .env and wrangler config files which often contain credentials. The script scripts/check-config.sh includes network functionality via curl, which could be used as a channel for data exfiltration if the agent is compromised.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 8, 2026, 07:30 PM