The Agent Skills Directory

[COMMAND_EXECUTION]: Multiple Python scripts, including 'render_resume_pdf.py', 'render_cover_letter_pdf.py', 'verify_resume_pdf.py', and 'update_application_tracker.py', utilize 'subprocess.run' to execute external binaries. These include LaTeX compilers such as 'latexmk', 'pdflatex', and 'tectonic', as well as the macOS system utility 'mdls' and other internal Python scripts. These commands are used to compile generated LaTeX source files into PDFs and manage the application workflow.
[DATA_EXFILTRATION]: The 'notion_sync.py' script transmits application tracking data (company, role, status, fit score, etc.) to the Notion API ('api.notion.com'). This process uses an integration token stored in the 'NOTION_TOKEN' environment variable. This behavior is a documented feature of the skill for keeping external trackers in sync.
[PROMPT_INJECTION]: The skill is designed to ingest and process untrusted data from external job description URLs or text provided by the user, as described in 'SKILL.md'. This data directly influences the content of tailored LaTeX resume files.
Ingestion points: External job description URLs and pasted text in the 'SKILL.md' workflow.
Boundary markers: The instructions lack explicit delimiting or instructions to ignore embedded commands within the job descriptions.
Capability inventory: The skill has the ability to write to the local filesystem, perform network requests to Notion, and execute shell commands via 'subprocess.run'.
Sanitization: A 'latex_escape' function is implemented in 'create_cover_letter.py' to sanitize specific user-provided strings. For the main resume content, the skill relies on the AI agent's own output filtering when editing the 'resume.tex' template.

resume-tailor