The Agent Skills Directory

[SAFE]: No malicious patterns, obfuscation, or unauthorized access attempts were detected across the skill files.
[COMMAND_EXECUTION]: The skill generates shell scripts using heredocs and executes them via the Bash tool. These scripts use proper quoting for variables (e.g., "$INPUT_FILE") and leverage standard tools (asciinema, jq, sed) to process data safely, mitigating command injection risks even with filenames containing spaces or special characters.
[DATA_EXFILTRATION]: The skill does not perform any network operations. Data ingestion and output are strictly limited to the local filesystem, as evidenced by the tool usage and code in workflow-phases.md and batch-workflow.md.
[PROMPT_INJECTION]: The skill includes an indirect prompt injection surface as it processes external recording files.
Ingestion points: Reads .cast files (NDJSON) in references/workflow-phases.md and references/batch-workflow.md.
Boundary markers: The skill does not explicitly use markers, but its primary function is to sanitize recordings by stripping ANSI codes.
Capability inventory: Uses Bash and Write tools to perform processing.
Sanitization: Employs jq for structured data extraction and sed patterns to remove potentially disruptive ANSI escape sequences.

asciinema-converter