The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses several powerful system commands including osascript to quit applications, pkill and kill -9 to terminate processes by name or PID, and open -a to restart the IDE. It also utilizes sqlite3 to query local data and mkdir to manage log directories.
[INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted data from a local SQLite database and system process listings to generate reports, creating a vulnerability surface.
Ingestion points: The skill reads data from cursor-metrics.sqlite3 and the output of the ps command as defined in SKILL.md.
Boundary markers: No explicit delimiters or boundary instructions are used when interpolating this external data into the agent's summary report.
Capability inventory: The skill possesses capabilities to terminate system processes (pkill, kill), execute UI scripts (osascript), and launch applications (open).
Sanitization: There is no evidence of sanitization or validation of the data retrieved from the SQLite database before it is ingested into the agent context.

nuke-cursor-app