The Agent Skills Directory

[SAFE]: The skill provides instructional templates and rhetorical strategies for academic writing. It does not contain any malicious code, prompt injections, or unauthorized data access patterns. All operations, including reading project configurations and writing drafts, are performed within the scope of the user's research project folder.
[DATA_EXPOSURE]: The skill interacts with local project files such as project.yaml and progress.yaml to manage the writing workflow. It does not attempt to access sensitive system files, environment variables, or hardcoded credentials.
[INDIRECT_PROMPT_INJECTION]: The skill includes an ingestion surface where it reads from project.yaml and processes user-provided drafts. While it has file-writing and git-execution capabilities, the risk is mitigated by the static nature of the commands and the specific intended use case of academic drafting.
[COMMAND_EXECUTION]: The skill instructions include standard git commands (git add and git commit) for version control of the generated drafts. These commands use static commit messages and are intended to manage the skill's own output files, posing no security risk to the host environment.

case-justification