The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests data from untrusted or external sources and interpolates them into generated configuration files, creating a surface for indirect prompt injection.
Ingestion points: Project manifests (package.json, pyproject.toml, Makefile), existing docs/dev-agents/config.md, and freeform user input (e.g., descriptions for 'Other' issue trackers).
Boundary markers: None explicitly mentioned for the interpolation of user-provided descriptions or manifest values into the generated config.md file.
Capability inventory: Executes shell commands via git and gh CLIs; writes and modifies files in the local repository structure (e.g., docs/dev-agents/, AGENTS.md, CLAUDE.md).
Sanitization: The skill instructions focus on idempotency and user confirmation before overwriting files but do not specify sanitization or escaping of the ingested data before it is written into configuration documents.
[COMMAND_EXECUTION]: The skill performs several command-line operations to detect environment state and configure the repository.
Evidence: The skill executes git rev-parse, git remote -v, git init, gh auth status, gh repo view, gh repo create, and gh label create to manage repository settings and metadata.

wkc-setup