Security audit
Security checks for vulnerabilities and agentic risk
This skill is for Splitwise expense creation, but it can create real account records with a long-lived token and gives agents broad automatic-invocation guidance without enough guardrails.
Review before installing. Only use this if you are comfortable giving the agent a long-lived Splitwise token and allowing it to create expense records. Keep the token out of prompts, logs, screenshots, and source files, and require your agent workflow to confirm payer, participant, amount, group, and description before submission.
The skill description overstates its capabilities by claiming balance checking and general expense management, while the implementation reportedly only creates a specific two-person 50/50 expense. This mismatch is dangerous because agents or users may authorize the skill under false assumptions, leading to mis-invocation of a write-capable action and reduced human oversight for financial operations.
The README instructs users to configure a long-lived Splitwise API token but provides no warning that this credential is highly sensitive, persistent, and should be stored securely. In a skill meant for agent use, this omission increases the chance users will expose the token in shell history, logs, screenshots, checked-in config files, or shared environments, enabling unauthorized access to Splitwise account data and actions.
The skill declares access to environment secrets and makes outbound network requests, but it does not define an explicit tool scope such as permissions or allowed-tools. That omission can cause an agent framework to grant broader capabilities than reviewers expect, reducing containment and making unintended secret access or network use harder to govern.
The autonomous invocation note gives a broad natural-language trigger for running a financial action without requiring explicit confirmation, parameter validation, or clear preconditions. In an agent setting, this can cause the skill to be invoked on ambiguous user requests and create real expenses using a long-lived token, making accidental or unauthorized transactions more likely.
No suspicious patterns detected.
