The Agent Skills Directory

[COMMAND_EXECUTION]: The scripts/changelog_generator.py script invokes the git binary to retrieve commit subjects. It utilizes subprocess.run with a list of arguments and shell=False, which is a security best practice for preventing command injection vulnerabilities. The arguments are strictly scoped to log retrieval.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes git commit logs, which are external and potentially untrusted data sources. If a commit message contains malicious instructions, the agent may attempt to follow them during the document generation process.\n
Ingestion points: Commit subjects are ingested from the local git history via scripts/changelog_generator.py and used as context for the agent.\n
Boundary markers: The script organizes output with Markdown headers (e.g., ## Features), which provides structure but does not guarantee the agent will ignore embedded instructions within the text.\n
Capability inventory: The skill is focused on text generation and roadmap structuring; however, the agent's wider capability set could be leveraged by injected instructions.\n
Sanitization: The tool performs no filtering, validation, or escaping of the commit message content before it is processed by the agent.

roadmap-communicator