The Agent Skills Directory

[COMMAND_EXECUTION]: The script roadmap-communicator/scripts/changelog_generator.py utilizes subprocess.run to execute the git log command.
Evidence: subprocess.run(cmd, capture_output=True, text=True, timeout=30) where cmd is ["git", "log", ...] in roadmap-communicator/scripts/changelog_generator.py.
Context: This command is used to extract commit subjects from the local repository to generate a structured changelog, which is the primary intended purpose of the script.
[PROMPT_INJECTION]: The skill set includes tools designed to parse and analyze untrusted external data, creating an indirect prompt injection surface.
Ingestion points: product-manager-toolkit/scripts/customer_interview_analyzer.py (analyzes interview transcripts), research-summarizer/scripts/extract_citations.py (parses academic papers and documents), and roadmap-communicator/scripts/changelog_generator.py (parses git commit messages).
Boundary markers: Absent. The tools output analysis results directly without using specific delimiters or instructions to the agent to disregard embedded content.
Capability inventory: Local file system read/write operations and execution of the git log command via subprocess.
Sanitization: None detected; the tools process text content using standard regex and string manipulation.
Context: This ingestion surface is a fundamental requirement for the skill's document analysis and summarization capabilities and represents a standard architectural risk for such tools.

product-skills