The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to run local validation commands using grep and perl to check the generated README for style compliance (e.g., detecting em dashes, improper headings, or relative image paths). These operations are restricted to the local README.md file.
[EXTERNAL_DOWNLOADS]: The skill verifies package existence and versions by querying official registries using npm view, cargo search, and pip index. These network operations target well-known, trusted services to ensure the accuracy of the generated documentation.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from untrusted project manifest files (e.g., package.json, Cargo.toml) and existing README files to extract metadata.
Ingestion points: Manifest files and existing project documentation.
Boundary markers: Not explicitly defined in the prompt instructions.
Capability inventory: Shell command execution (grep, perl, npm, cargo, pip) and file system write access.
Sanitization: No specific sanitization or escaping of extracted metadata is defined before it is used in the workflow.

readme-creator