The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to use the rg (ripgrep) utility to perform project-wide searches for specific Java annotations and file structures (e.g., searching for @RestController, @Service, or @Mapper). This is a standard and necessary function for the skill's stated purpose of source code analysis.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and prioritizes project documentation and existing AI-generated files that could contain malicious instructions designed to manipulate the agent's behavior.
Ingestion points: The agent is directed to read files from paths such as docs/**/*.md, docs/superpowers/, and graphify-out/ as specified in references/ai-doc-discovery.md.
Boundary markers: Present. The skill implements 'Hard Rules' in SKILL.md that define source code as the absolute authority and explicitly flag generated documentation as 'drafts' until reviewed by a team lead or architect.
Capability inventory: The skill has the capability to read project files, write new documentation files to the local disk, and execute shell commands (rg) as described in references/source-scan-guide.md.
Sanitization: Absent. While the skill classifies ingested documents into categories like factual-context and prompt-template, it does not perform automated sanitization or filtering of the text content to prevent the execution of embedded instructions.

project-context-init