The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from repository documentation and the codebase to curate an onboarding guide. It lacks safeguards to prevent malicious instructions embedded in these source files from being included in the final output.
Ingestion points: Documentation files in the .chalk/docs/ directory tree and project configuration files (e.g., package.json, .env.example).
Boundary markers: No specific delimiters or "ignore instructions" warnings are utilized when processing external documentation.
Capability inventory: The skill uses Read, Glob, Grep, Bash, and Write tools across its workflow.
Sanitization: No sanitization or validation logic is defined for the content extracted from the project files.
[COMMAND_EXECUTION]: The skill employs the Bash tool to inspect project structures, dependency managers, and build scripts. While intended for reconnaissance to populate the onboarding guide, this capability could be exploited if the agent's logic is subverted through indirect injection.

create-onboarding-guide