The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell commands via Bash to handle large file writes and execute the LaTeX compilation toolchain (latexmk). It specifically instructs the agent to retry writes using cat and to do so silently without asking for user permission.
[PROMPT_INJECTION]: The skill processes user-provided narrative reports and research data which presents a surface for indirect prompt injection. 1. Ingestion points: Reads content from NARRATIVE_REPORT.md and other research data files. 2. Boundary markers: No explicit delimiters are used to separate user-provided data from system instructions. 3. Capability inventory: The skill has access to Bash, Write, and Agent tools, allowing for filesystem modification and further task delegation. 4. Sanitization: No sanitization or verification of the input files is described.
[DATA_EXFILTRATION]: The skill transmits research narrative and paper content to an external model (gpt-5.4) via the Codex MCP for the purpose of review and automated improvement rounds.

paper-writing