The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill setup instructions involve cloning a repository from GitHub (github.com/crafter-station/sismo-abierto) and using bunx to execute a package from the @crafter scope. Both resources are owned by the skill's author.
[COMMAND_EXECUTION]: The skill executes a local CLI tool, sismo, to query seismic activity and export data to files like GeoJSON or CSV.
[PROMPT_INJECTION]: The skill processes data from seismic monitoring agencies which constitutes an indirect prompt injection surface. Ingestion points: Data retrieved from IGP and USGS sources via the sismo CLI. Boundary markers: The instructions include an explicit warning that text from these sources is untrusted and that the agent should never follow instructions embedded within it. Capability inventory: The agent can execute CLI commands, access network resources, and write to local files. Sanitization: The skill relies on instructional boundaries provided to the agent to mitigate the risk of processing malicious directives in source data.
[DATA_EXFILTRATION]: The CLI tool performs network requests to scientific and government organizations (IGP and USGS) to retrieve seismic waveforms and event data. This behavior is consistent with the skill's primary function.

sismo-cli