The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes the evolver CLI and environment verification commands such as node, git, and npm install. It also runs a local helper script ~/plugins/evolver/scripts/evolver-status.js to check workspace health.
[PERSISTENCE]: The skill modifies agent configuration and hooks in the user's home directory, specifically updating ~/.codex/AGENTS.md and injecting platform hooks to maintain integration across sessions.
[INDIRECT_PROMPT_INJECTION]:
Ingestion points: External data is ingested through the evolver_recipe_search, evolver_search_assets, and evolver_poll tools which read from the Evolver Hub as described in SKILL.md.
Boundary markers: The engine utilizes "protocol-bound GEP prompts" and generates EvolutionEvents to provide an audit trail for engine outputs.
Capability inventory: The skill possesses capabilities for package installation, home directory file modification, and network communication with evomap.ai.
Sanitization: A local Evolver Proxy is employed to gate content quality and manage local outcomes before they are processed by the agent.
[EXTERNAL_DOWNLOADS]: The skill fetches the official @evomap/evolver package from the npm registry and interacts with the vendor's Hub infrastructure at evomap.ai for asset discovery and publishing.

evolver