The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface through its logging implementation.
Ingestion points: references/monitoring.md demonstrates logging of args.name and error.message from tool calls and runtime errors.
Boundary markers: The implementation does not use explicit boundary markers to isolate these untrusted inputs within the logs.
Capability inventory: Logs are written to stderr or local log files as configured in the LOG_DESTINATION and LOG_FILE environment variables defined in SKILL.md.
Sanitization: There is no evidence of sanitization or escaping of external data before it is recorded in the structured logs.
[EXTERNAL_DOWNLOADS]: The skill involves standard package management and connections to external service providers.
Downloads Node.js packages using bun install during the Docker build process in assets/Dockerfile.
Utilizes external Nostr relays for communication, including well-known public relays like wss://nos.lol and vendor-managed infrastructure at wss://relay.contextvm.org.

deployment