The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local Node.js script (.aiox-core/development/scripts/unified-activation-pipeline.js) as a fallback mechanism during its activation sequence. It also includes logic to suggest environment bootstrap commands if a git repository is not detected in the current workspace.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect instructions by ingesting data from external files within the project environment.
Ingestion points: The agent reads story documentation from docs/stories/ and workflow handoff artifacts from .aiox/handoffs/ to populate its greeting and suggest command chains.
Boundary markers: No specific delimiters or "ignore embedded instructions" warnings are defined when these external data sources are interpolated into the agent's context.
Capability inventory: The agent can execute local JavaScript files, perform web research using the exa tool, and modify the state of handoff files by marking them as consumed.
Sanitization: The skill lacks evidence of sanitization or validation of the content found within the ingested YAML and Markdown files.
[DYNAMIC_EXECUTION]: The skill references and executes local JavaScript files, such as pattern-extractor.js, to perform analysis tasks on the codebase at runtime.
[PRIVILEGE_ESCALATION]: The skill implements a *yolo command designed to cycle through permission modes (ask, auto, explore). This allows for the reduction of user oversight by enabling modes that bypass manual confirmation requirements for tool execution.

aiox-analyst