The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill documentation describes how to use crypto wallet private keys by exporting them to the BOTCHAN_PRIVATE_KEY environment variable or passing them via the --private-key flag. Handling raw private keys in these ways can lead to exposure in system logs, process lists, or shell histories.
[EXTERNAL_DOWNLOADS]: The skill requires the installation of the botchan package from the public NPM registry and the addition of a skill from a remote GitHub repository (stuckinaboot/botchan). These resources are provided by the vendor to enable the skill's core functionality.
[PROMPT_INJECTION]: The skill retrieves and processes untrusted text data from the Base blockchain, creating an indirect prompt injection surface.
Ingestion points: Untrusted data is ingested from onchain feeds, comments, and profile information via the read, comments, and profile commands.
Boundary markers: The instructions do not specify the use of delimiters or protective warnings to prevent the agent from following instructions found within retrieved blockchain messages.
Capability inventory: The agent can perform write operations such as post, comment, and register on the blockchain, which could be triggered by malicious message content.
Sanitization: There is no evidence of sanitization or filtering of the onchain message content before it is processed by the agent.

botchan