The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: Fetches and installs additional skills from the vendor's official GitHub repository and the AgentMail repository using the npx skills add command to provide extended functionality.
[COMMAND_EXECUTION]: Executes shell commands to detect the runtime environment, manage local configuration files, and generate a maintenance script for registration health checks.
[DATA_EXFILTRATION]: Facilitates account registration and session management by communicating with the resolved.sh and agentmail.to APIs. This involves sending the agent's email address and retrieving session tokens and API keys.
[CREDENTIALS_UNSAFE]: Manages sensitive information such as API keys and session tokens by temporarily storing them in the /tmp directory. This data is used to automate account creation and resource registration without human intervention.
[PROMPT_INJECTION]: Processes external data from email messages to extract authentication tokens, creating a surface for indirect instructions.
Ingestion points: Retrieves message content from the AgentMail API within the token polling phase in SKILL.md.
Boundary markers: Absent; the skill uses regex to locate specific token patterns within the message body.
Capability inventory: Uses curl for network requests and python3 for string parsing and file writing.
Sanitization: Employs regex to isolate the token parameter, which restricts the processed input to expected alphanumeric characters.

rstack-bootstrap