The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process a high volume of untrusted data from the local repository, including documentation files, source code, and configuration files, to reconstruct the documentation system.
Ingestion points: Steps 1, 2, and 5 in SKILL.md describe reading agent entrypoints (e.g., AGENTS.md, CLAUDE.md), README files, code, and previously moved documentation.
Boundary markers: The skill describes an 'Instruction Authority' mechanism to resolve rule conflicts but does not explicitly mention using delimiters or specific isolation techniques to prevent the agent from following instructions embedded within the files it harvests.
Capability inventory: The skill moves files to a backup directory, writes new documentation files, and prunes stale files/directories using repository tools.
Sanitization: No explicit content sanitization or escaping mechanisms are defined for the repository content interpolated into the agent's context.
[COMMAND_EXECUTION]: The skill invokes shell commands and repository-specific tools to manage state and validate documentation.
Git Operations: The workflow uses git status --short and git diff --check to monitor the working tree state.
Local Tooling: Step 6 involves executing 'docs/skill validators available in the repo' and 'docs-as-code generators,' which may trigger the execution of arbitrary scripts or binaries present in the local environment.

documentation-reseed