The Agent Skills Directory

[COMMAND_EXECUTION] (LOW): The script safe_pack.py executes the external command repomix using subprocess.run to perform the packaging step. This is a standard part of the skill's workflow and is executed without shell=True, minimizing common injection vulnerabilities.
[PROMPT_INJECTION] (LOW): The skill is susceptible to indirect prompt injection because it reads and displays content from untrusted files in a user-provided directory.
Ingestion points: scan_secrets.py (line 77) reads files from the target directory.
Boundary markers: Absent; the script prints raw matches and line contexts directly to stdout.
Capability inventory: Access to subprocess.run (executing repomix) and local file reading via open.
Sanitization: Absent; though output is truncated for length, it is not sanitized or escaped to prevent an LLM from interpreting instructions contained within the scanned files.

repomix-safe-mixer