The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill instructions in commands/docs-claude.md and commands/docs-update.md specify reading from sensitive configuration files, including .env and .env.local, to parse environment variable names for documentation auditing purposes.
[INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection attack surface by ingesting data from untrusted project files to populate documentation templates.
Ingestion points: Reads metadata and configuration from package.json, wrangler.jsonc, .env files, and git remote URLs.
Boundary markers: The skill uses {{PLACEHOLDER}} syntax for interpolation within markdown templates but does not implement explicit delimiters or instructions to ignore embedded commands in the source data.
Capability inventory: The skill possesses file system modification capabilities including mkdir, mv, rm (deletion), and file writing.
Sanitization: There is no evidence of sanitization, validation, or escaping of the ingested data before it is written to the documentation files.
[COMMAND_EXECUTION]: The skill workflow involves the execution of shell commands for file management, such as mkdir for directory scaffolding, mv for archiving documentation, and rm for deleting orphaned files as described in the docs-update command process.
[EXTERNAL_DOWNLOADS]: The skill performs network operations to the well-known npm registry using npm view to check for package version updates during the documentation audit process.

docs-workflow