The Agent Skills Directory

[PROMPT_INJECTION]: The skill instructs the agent to configure a 'YOLO' mode within the project's documentation (AGENTS.md and CLAUDE.md). This mode is explicitly described as granting 'standing authorization' and bypassing 'generic prompt confirmation requirements.' This acts as a programmed instruction to override typical AI safety guardrails and user confirmation loops.
[DYNAMIC_EXECUTION]: The skill performs runtime execution of local Python scripts, specifically detect-architecture.py and migrate-workspace.py, which are invoked with project-level arguments.
[COMMAND_EXECUTION]: The skill uses shell commands to run Python scripts (python3 <plugin-root>/scripts/detect-architecture.py) to inspect the host project environment.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and merge data from various files in the local repository, which could be influenced by external actors if the repository content is untrusted.
Ingestion points: The skill reads AGENTS.md, CLAUDE.md, .lovable-agent/config.json, and .lovable-agent/context.md from the project root.
Boundary markers: No explicit delimiters or boundary markers are defined to separate user-provided content from agent instructions during the merge process.
Capability inventory: The skill has capabilities to read/write files and execute shell-based Python scripts.
Sanitization: The instructions do not specify any validation or sanitization steps for the prose or configuration fields imported from the repository files.

init