The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the asciinema CLI to capture terminal output. This is the core functionality and is used to record either interactive sessions or specific command strings.
[DATA_EXFILTRATION]: The instructions include extensive security best practices to prevent sensitive data from being captured in recordings, such as using temporary home directories (mktemp -d) and stripped environments (env -i) to hide host details, SSH keys, or cloud credentials.
[REMOTE_CODE_EXECUTION]: The skill provides instructions for recording the execution of local scripts (e.g., via the -c flag). It explicitly advises users to review these scripts before recording, treating them as executable code to avoid unintended actions.
[SAFE]: The skill was evaluated for indirect prompt injection risks associated with processing terminal output: 1. Ingestion points: terminal session content captured in .cast files (SKILL.md); 2. Boundary markers: specific recommendations for isolated environments and generic prompts; 3. Capability inventory: asciinema recording and playback; 4. Sanitization: warnings to perform secret scans (cast-redact) before publishing recordings.

cast-record