The Agent Skills Directory

Prompt Injection (LOW): The skill is susceptible to indirect prompt injection because it ingests untrusted data from git commits, branch changes, and file diffs.
Ingestion points: Processes data from the git root, including commits, branch changes, and file diffs as described in the PR Description Generation section.
Boundary markers: No explicit delimiters or instructions to ignore embedded commands are specified in the framework to separate untrusted data from system instructions.
Capability inventory: Utilizes tools such as memory_search_all, pm_portfolio, and pm_ask for data retrieval and analysis across the codebase ecosystem.
Sanitization: No sanitization or escaping of external content is mentioned for the interpolated data during description generation.

portfolio-intelligence