The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell commands and Python sub-processes to perform local system discovery. It executes git, find, and ls to audit directories and retrieve commit history from local repositories, specifically looking for paths under ~/Code/github.com/Soul-Brews-Studio/.
[DATA_EXFILTRATION]: The skill collects and transmits metadata to external network addresses. It audits agent configuration files (e.g., .claude/settings*.json) to count installed skills and uses a custom CLI tool (maw hey) to send this inventory data along with repository status to remote nodes defined in a contacts.json file.
[PROMPT_INJECTION]: The skill has a surface for indirect prompt injection by ingesting untrusted data from external sources. It parses a local contacts.json file and processes responses from remote network pings. If these external sources contain malicious instructions, they could influence the agent's behavior when the reporting step is executed.
Ingestion points: Reads $PSI/contacts.json and captures output from maw hey network commands.
Boundary markers: No explicit delimiters or instructions are used to separate ingested metadata from the agent's execution instructions.
Capability inventory: The skill has the capability to execute shell commands (git, find), read system files, and perform network operations.
Sanitization: There is no evidence of sanitization or validation performed on the data retrieved from remote nodes before it is presented in the report.

fleet