The Agent Skills Directory

[SAFE]: The Python script scripts/changelog_generator.py executes git log using subprocess.run with a list of arguments, which is the secure way to execute external commands and prevents command injection.
[INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted commit messages from git logs. While this creates a surface for indirect prompt injection, it is inherent to the skill's purpose and does not present a direct threat.
Ingestion points: The scripts/changelog_generator.py script reads commit subjects from git log or stdin.
Boundary markers: Output is formatted into Markdown sections, but lacks explicit security delimiters for the ingested content.
Capability inventory: The skill uses subprocess calls to read git logs and provides templates for the agent.
Sanitization: The script organizes commit messages but does not sanitize their contents for potential injection strings.

roadmap-communicator