The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The installation section contains the command curl -sSL https://raw.githubusercontent.com/fiatjaf/nak/master/install.sh | sh. This pattern downloads a shell script from a remote GitHub repository and pipes it directly into the shell for execution, allowing for arbitrary code execution on the user's system without prior verification.
[COMMAND_EXECUTION]: The skill provides numerous examples of shell commands for managing Nostr identities and publishing events. Some examples involve handling sensitive private keys, such as setting the NOSTR_SECRET_KEY environment variable or reading keys from local files using subshells (e.g., --sec $(cat /path/to/key.txt)).
[INDIRECT_PROMPT_INJECTION]: The skill utilizes functionality to fetch data from external and untrusted Nostr relays, which is a significant surface for indirect prompt injection attacks.
Ingestion points: Data enters the agent's context through nak fetch and nak req commands documented in SKILL.md.
Boundary markers: There are no instructions or delimiters provided to warn the agent about potential instructions embedded in the fetched Nostr events.
Capability inventory: The skill enables high-impact capabilities including key management, event signing, and publishing to network relays.
Sanitization: No sanitization or validation of the fetched event content is specified, meaning malicious events containing instructions could potentially manipulate the agent's actions.

nak