fable-prove-it
Pass
Audited by Gen Agent Trust Hub on Oct 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions create a surface for indirect prompt injection by processing external user-provided data while the agent has access to terminal commands.
- Ingestion points: The skill ingests untrusted data through the
candidate_claiminput field defined inSKILL.md. - Boundary markers: The instructions do not specify the use of delimiters or 'ignore' instructions to prevent the agent from following commands potentially embedded within the
candidate_claim. - Capability inventory: The skill requires
terminal_accessand explicitly directs the agent to execute shell commands to verify claims, which could be exploited if an attacker-controlled claim influences the command choice. - Sanitization: There are no documented sanitization, validation, or escaping procedures for the content of the
candidate_claimbefore it is used to guide agent actions.
Audit Metadata