The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local Python packages and modules via uv run python, specifically the cli_council package for multi-model verification results.
[COMMAND_EXECUTION]: Fallback verification logic in references/openalex-verification.md uses uv run python -c to execute Python code strings and dynamically modifies the sys.path to import local modules from the .scripts/openalex directory.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests and processes untrusted content from .tex and .bib files.
Ingestion points: Scans .tex files for citations and parses @entry keys from .bib files (SKILL.md).
Boundary markers: None present; the skill lacks delimiters or instructions to ignore embedded agent commands within the bibliography data.
Capability inventory: The skill has permissions for Bash operations (ls, mkdir, rm), Write operations, and uv run python execution.
Sanitization: There is no documented sanitization or validation of the extracted citation metadata before it is processed by the agent or sub-agents.

validate-bib