The Agent Skills Directory

[COMMAND_EXECUTION]: The skill performs blockchain transactions using callContract in stackspot.ts to interact with smart contracts on the Stacks mainnet.- [DATA_EXFILTRATION]: All write commands in stackspot.ts (lines 255, 283, 313, 342) use PostConditionMode.Allow. This setting bypasses standard security checks that protect wallet assets, allowing the target contracts to move any tokens from the user's account during the transaction. This is a high-risk configuration that increases exposure if the destination contracts are compromised.- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface.
Ingestion points: Data is retrieved from on-chain contracts via callPotReadOnly in stackspot.ts.
Boundary markers: There are no protective markers or instructions used when outputting contract data to the agent context.
Capability inventory: The skill can execute financial transactions and manage wallet states.
Sanitization: Retrieved data is not sanitized before being presented to the agent, potentially allowing malicious on-chain content to influence agent logic.

stackspot