The Agent Skills Directory

[COMMAND_EXECUTION]: The skill invokes shell commands using uv run to execute scripts located in a separate skill's directory. Specifically, it calls openalex_cli.py, europepmc_api.py, and read_paper.py located within the literature-review-dir.
[DYNAMIC_EXECUTION]: The skill executes scripts from a computed or placeholder path (<literature-review-dir>), indicating a functional dependency on the file structure of another skill.
[INDIRECT_PROMPT_INJECTION]:
Ingestion points: The skill ingests untrusted data from the open web, news sites, and social media platforms using the WebSearch and WebFetch tools.
Boundary markers: The instructions do not define explicit boundary markers or delimiters (e.g., XML tags or triple backticks) to isolate external content from the agent's internal reasoning context.
Capability inventory: The skill has access to the Bash tool and performs script execution via uv run, which could be targeted by instructions embedded in malicious research sources.
Sanitization: There are no instructions for sanitizing or validating the content retrieved from external sources before it is processed or used to generate investigative reports.

multi-source-investigation