The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes standard system commands including mkdir, cp, mv, ln, and rmdir to manage the project's context store and symlinks. These operations are restricted to the repository root and a project-specific directory within ~/.codevoyant/.
[DYNAMIC_EXECUTION]: The skill utilizes local Python scripts and inline Python execution via heredocs for processing JSON metadata and normalizing Markdown registries. These scripts use standard library modules (json, sys, os, re) and do not process untrusted code from external sources.
[DATA_EXPOSURE]: The skill manages project context data stored in ~/.codevoyant/. This directory is used cross-worktree to maintain project state and does not involve accessing sensitive system credentials or personal user data.
[INDIRECT_PROMPT_INJECTION]: The migration logic ingests data from local store files (metadata.json, README.md) and plan directories.
Ingestion points: Reads version tags from migration files and status records from the store registry.
Boundary markers: Commands are gated by a manual confirmation step requiring the user to approve the specific list of migration files before execution.
Capability inventory: Limited to file management and structured data manipulation within the scoped project store.
Sanitization: Uses standard JSON and string parsing to handle store content.

migrate