The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted market data from an external API, which could potentially be used to influence agent behavior through embedded instructions.
Ingestion points: Market titles, descriptions, and metadata are fetched from api.stacksmarket.app via the list-markets, search-markets, and get-market subcommands in stacks-market.ts.
Boundary markers: No explicit boundary markers or specific instructions to ignore embedded content are used when passing retrieved market data to the agent context.
Capability inventory: The skill possesses write capabilities to the Stacks blockchain, including buying/selling shares and redeeming funds via smart contract calls in stacks-market.ts.
Sanitization: There is no evidence of content sanitization or validation for the strings retrieved from the external market API.
[COMMAND_EXECUTION]: Executes on-chain transactions that move funds on the Stacks blockchain.
Evidence: The buy-yes, buy-no, sell-yes, sell-no, and redeem commands in stacks-market.ts use the callContract function to perform state-changing operations on the market-factory-v18-bias contract.
The skill uses PostConditionMode.Allow, which permits the contract to transfer assets without the granular restrictions typically provided by explicit Stacks post-conditions; this is noted as a requirement for this specific contract's internal logic.
[EXTERNAL_DOWNLOADS]: Fetches data from an external service to facilitate market discovery.
Evidence: The fetchMarketApi function in stacks-market.ts makes network requests to https://api.stacksmarket.app to retrieve prediction market listings and details.

stacks-market