The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from various document formats which could contain malicious instructions designed to subvert the agent's behavior.
Ingestion points: SKILL.md Phase 2.1 uses the Glob and Read tools to scan and read user-provided materials (.pdf, .pptx, .md, .txt, .png, .jpg).
Boundary markers: The skill does not define clear delimiters or specific instructions to the agent to ignore potentially malicious embedded content within the materials.
Capability inventory: The agent has access to the Bash tool (used for parsing and LaTeX compilation) and the Write tool (used to generate the output files).
Sanitization: No explicit sanitization or filtering is performed on the content extracted from the files before it is processed by the agent.
[COMMAND_EXECUTION]: The skill uses the Bash tool to execute Python commands and custom scripts for extracting text and images from documents, and for compiling LaTeX output.
Evidence: SKILL.md contains multiple python -c snippets to parse PDFs and PPTX files. It also executes scripts/config_server.py and scripts/editor_server.py to facilitate user interaction.
[DYNAMIC_EXECUTION]: The skill generates LaTeX source code at runtime by interpolating user-provided configuration values and extracted material content into templates.
Evidence: Phase 2.3 of SKILL.md describes the assembly of cheatsheet.tex using placeholders in templates/cheatsheet_base.tex. The resulting file is subsequently processed by local LaTeX tools via scripts/editor_server.py.

cheatsheet-generator