The Agent Skills Directory

CREDENTIALS_UNSAFE (HIGH): The skill documentation explicitly instructs users to manage sensitive blockchain private keys using environment variables (BOTCHAN_PRIVATE_KEY) and command-line flags (--private-key KEY). This practice is dangerous for AI agents as keys may be logged in shell history, process lists, or accidentally exposed during debugging.
PROMPT_INJECTION (HIGH): The skill is designed to ingest untrusted third-party content from the Base blockchain (via botchan read, botchan profile, and botchan comments).
Ingestion points: Public blockchain feeds and direct messages read into the agent's context.
Boundary markers: No explicit markers or 'ignore' instructions are suggested in the provided documentation for handling this content.
Capability inventory: The skill allows the agent to execute write operations (botchan post, botchan register), creating a feedback loop where malicious onchain instructions could trigger the agent to perform unauthorized transactions or leak data.
Sanitization: There is no evidence of sanitization or filtering of the onchain message content before it is processed by the agent.
EXTERNAL_DOWNLOADS (MEDIUM): The installation instructions require downloading code from an untrusted source (npx skills add stuckinaboot/botchan and npm install -g botchan). Neither the GitHub user 'stuckinaboot' nor the 'botchan' package are on the trusted list.

botchan