The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is configured to execute a local shell script ./scripts/session-control.sh to update session state files in the .claude/state/ directory.
[PROMPT_INJECTION]: The skill processes workflow variables by interpolating them into bash command strings, establishing a surface for indirect prompt injection where malicious input could alter the command's execution.
Ingestion points: External variables resume_session_id, resume_latest, fork_session_id, and fork_reason are used as CLI arguments in references/session-control.md.
Boundary markers: Absent. The skill does not employ delimiters or instructions to ignore potential commands embedded within the variable data.
Capability inventory: Use of the Bash tool for script execution and Write/Edit tools for modifying state files.
Sanitization: Absent. The documentation does not specify any escaping or validation mechanisms for the interpolated workflow variables.

session-control