The Agent Skills Directory

[PROMPT_INJECTION]: The skill presents an indirect prompt injection surface by ingesting untrusted data (bounty titles, descriptions, and tags) from a remote API and displaying it directly to the agent.
Ingestion points: Data enters the agent's context via the scan, match, and detail commands in bounty-scanner.ts which fetch data from https://bounty.drx4.xyz/api.
Boundary markers: None. The external data is interpolated into JSON output without delimiters or warnings to ignore embedded instructions.
Capability inventory: The agent is empowered to perform automated claims and generate cryptographic signatures using the signing skill.
Sanitization: No sanitization or safety filtering is performed on the bounty content before it is processed by the matching logic or presented to the agent.
[EXTERNAL_DOWNLOADS]: The skill makes frequent network requests to https://bounty.drx4.xyz/api to sync bounty data and statistics. This domain is the primary data source for the skill's functionality.
[DATA_EXFILTRATION]: The claim flow instructs the agent to send sensitive information, including a BTC address and a cryptographic signature, to a dynamic endpoint provided by the API. While essential for the skill's purpose, this pattern poses a risk if the API returns a malicious or third-party endpoint to harvest signatures.

bounty-scanner