cpsc-recall-check
Pass
Audited by Gen Agent Trust Hub on Aug 18, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONCREDENTIALS_UNSAFEREMOTE_CODE_EXECUTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill instructions direct the agent to fetch a script (
call.mjs) from the vendor's domain athttps://fittings.sh/skill/call.mjsfor local execution. This script is used to interact with the API and manage payment handshakes. - [COMMAND_EXECUTION]: The agent is expected to run the
node call.mjscommand to perform recall checks. This script handles network requests and cryptographic signature generation. - [DATA_EXPOSURE_AND_EXFILTRATION]: The skill utilizes a private key stored in the
FITTINGS_PRIVATE_KEYenvironment variable. This key is used by the execution script to sign payment authorizations. While the documentation provides safety guidance regarding throwaway wallets, the key is processed in plaintext by the agent. - [INDIRECT_PROMPT_INJECTION]: The skill ingests recall data from an external vendor-controlled API, creating a surface for indirect prompt injection.
- Ingestion points: Data is retrieved from
https://fittings.sh/v1/cpsc/recall-checkthrough JSON responses. - Boundary markers: No explicit markers or instructions are provided to the agent to distinguish external data from core instructions.
- Capability inventory: The agent has capabilities for shell execution, network access, and transaction signing.
- Sanitization: The script outputs raw JSON directly from the API without sanitization or validation of the content.
Audit Metadata