secrets

Pass

Audited by Gen Agent Trust Hub on Mar 9, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection via the processing of untrusted input in the form of connection strings.
  • Ingestion points: SKILL.md (Step 3) and scripts/parse-connection-string.sh accept and process arbitrary strings from users.
  • Boundary markers: No explicit boundary markers or instructions to disregard embedded commands in the input data are present.
  • Capability inventory: The skill executes shell scripts and CLI commands (goldsky secret create, goldsky secret list, goldsky secret reveal) which could be influenced if a user provides a maliciously crafted connection string that evades simple parsing.
  • Sanitization: The parse-connection-string.sh script performs basic string replacement but lacks robust validation to ensure inputs are strictly valid connection strings before being used in CLI commands.
  • [COMMAND_EXECUTION]: The skill executes a bundled shell script (scripts/parse-connection-string.sh) and the vendor's CLI tool (goldsky). These operations are required for the skill's stated purpose of secrets management and are executed locally.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 9, 2026, 11:33 PM