The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes a local Python script, scripts/rfc_index.py, to perform bookkeeping tasks such as allocating RFC numbers, instantiating templates, and validating index consistency. This script is part of the skill package and operates strictly on local project files within the rfcs/ or rfc/ directory using standard file I/O operations.
[DATA_EXPOSURE]: The skill reads and writes design documents and an index file in the repository's RFC directory. The analysis confirms that the skill does not access sensitive system paths (like ~/.ssh or .env files) or exfiltrate data to external locations.
[INDIRECT_PROMPT_INJECTION]: The skill processes and interpolates data from local RFC files. The risk of indirect injection is mitigated by the use of defined templates and sanitization logic within the rfc_index.py script, including title slugification to prevent malicious filenames and escaping of special characters in Markdown table cells.
Ingestion points: Reads rfcs/INDEX.md and rfcs/NNNN-*.md files from the local repository (scripts/rfc_index.py).
Boundary markers: The skill enforces structure through the use of specific Markdown templates (references/rfc-template.md and references/index-template.md) and numbered sections.
Capability inventory: Performs file read/write operations and shell execution of the included helper script.
Sanitization: Implements regex-based parsing, title slugification, and GFM table cell escaping to ensure content integrity.

rfc-writer