The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides an example command in SKILL.md to execute a local Python script (generate_openrouter_schematic.py) for figure generation. This execution is scoped to the primary purpose of the skill and uses standard environment variable configurations for API keys.
[EXTERNAL_DOWNLOADS]: The skill references an external dependency, the nature-figure skill, which must be installed separately. This is a common modular dependency pattern and targets a known functional extension for figure creation.
[INDIRECT_PROMPT_INJECTION]: The skill contains a vulnerability surface as it is designed to ingest and process untrusted external data.
Ingestion points: The skill gathers data from "verified local notes" and "trusted web sources" as part of its planning stage in SKILL.md.
Boundary markers: There are no explicit instructions or delimiters provided to the agent to distinguish between its own instructions and the content of the ingested data.
Capability inventory: The skill possesses the capability to execute shell commands (via python in SKILL.md) and write to the local file system (via the --outdir parameter).
Sanitization: No sanitization or validation of the ingested external content is specified before it is integrated into the manuscript drafting process.

scientific-writing