The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it is designed to ingest and process untrusted data from external repositories (source code, README files, and documentation) to generate summaries.
Ingestion points: The skill reads file content from local paths and remote GitHub repositories as defined in the '调查手順' (Investigation Steps) and 'コーパスを作る' (Creating Corpus) sections of SKILL.md.
Boundary markers: There are no explicit instructions to use XML tags or other delimiters to isolate external repository content from the agent's system instructions.
Capability inventory: The skill utilizes shell commands including rg (ripgrep) for searching, git for metadata retrieval, and mmdc (Mermaid CLI) for diagram generation. It also performs file system read and write operations within the user-defined output directory.
Sanitization: The instructions do not include specific sanitization, filtering, or validation steps for the content extracted from repositories before it is processed by the AI model.

extract-glossary