mine

Pass

Audited by Gen Agent Trust Hub on Jun 18, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill instructions end with a raw $ARGUMENTS interpolation. This allows user-provided text to be appended directly to the system prompt, potentially enabling an attacker to override the agent's behavior.
  • [PROMPT_INJECTION]: Indirect prompt injection vulnerability through the processing of untrusted session data.
  • Ingestion points: The skill reads and processes session JSONL files which contain external content, including human messages and tool outputs (SKILL.md).
  • Boundary markers: The skill does not implement boundary markers or instructions for the agent to ignore potentially malicious content within the session logs.
  • Capability inventory: The skill uses 4 parallel subagents to process the data and display assistant responses, tool calls, bash commands, and code snippets.
  • Sanitization: There is no evidence of sanitization, validation, or escaping of the content ingested from the session files before it is processed by the subagents.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 18, 2026, 01:15 PM
Security Audit — agent-trust-hub — mine