Aphorisms
Pass
Audited by Gen Agent Trust Hub on Jun 30, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection. Ingestion points: The skill fetches untrusted data via 'WebFetch' in 'Workflows/FindAphorism.md' and 'WebSearch' in 'Workflows/AddAphorism.md' and 'Workflows/ResearchThinker.md'. Boundary markers: The workflows do not specify any delimiters or instructions to the agent to ignore instructions embedded within the fetched content. Capability inventory: The skill environment allows for shell command execution (via curl and echo) and file system modification (via the Edit tool). Sanitization: There is no mention of filtering, escaping, or validating external content before it is processed for theme matching or summarization.
- [COMMAND_EXECUTION]: The skill directs the agent to execute shell commands for notifications and telemetry. The execution log in 'SKILL.md' instructs the agent to interpolate a user-provided input summary ('8_WORD_SUMMARY') into an 'echo' command. This pattern is vulnerable to command injection if the summary contains shell control characters and is not strictly sanitized by the agent before execution.
Audit Metadata