The Agent Skills Directory

[SAFE]: No malicious patterns such as prompt injection, obfuscation, or unauthorized remote code execution were found in the skill instructions.
[DATA_EXPOSURE]: The skill operates within the /data/boom/Papers directory. This specific file access is documented and aligns with the skill's purpose for project-specific manuscript management.
[PROMPT_INJECTION]: There are no instructions that attempt to override system prompts or bypass safety guidelines. The instructions focus on routing and sequencing tasks.
[INDIRECT_PROMPT_INJECTION]: The skill describes a workflow involving the reading and updating of local project files (e.g., project_truth.md, result_summary.md). While these represent ingestion points for external data into the agent's context, the skill itself provides purely instructional guidance and does not execute dangerous commands on the content.

paper-workflow