The Agent Skills Directory

[SAFE]: The skill follows standard development planning procedures without using network access or executing code. Actions are limited to reading project files and writing markdown plan documents to the docs/plans/ directory.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from the local filesystem to inform its generated plans. * Ingestion points: Relevant code, docs, and config files are read via Read, Grep, and Glob. * Boundary markers: No explicit delimiters are used to separate user instructions from file content. * Capability inventory: The agent can Write and Edit files and use AskUserQuestion. * Sanitization: No explicit content sanitization or validation is implemented.

plankit