The Agent Skills Directory

[SAFE]: The skill's operations are limited to project documentation management. It does not include hardcoded credentials, network communication, or administrative privilege requests.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from the project codebase and existing documentation, which creates a theoretical attack surface for indirect prompt injection.
Ingestion points: Reads source code files, GLOSSARY.md, GLOSSARY-MAP.md, and docs/adr/*.md.
Boundary markers: None identified; the agent processes these files directly to resolve terminology.
Capability inventory: The skill can read local files and write/update Markdown files. It lacks network access and arbitrary command execution.
Sanitization: No explicit sanitization of codebase content is performed before processing.
Risk Assessment: The behavior is considered safe because the skill only interacts with local documentation and lacks high-privilege capabilities that could be abused via injection.

domain-modeling