The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to use local tools such as grep and glob, and to execute a specific Python script at .claude/scripts/code_graph. These actions are integrated into the intended workflow for analyzing code structure and dependencies.
[PROMPT_INJECTION]: The skill uses highly assertive and repetitive language (e.g., "MANDATORY IMPORTANT MUST ATTENTION") to enforce a specific procedural "hard-gate." This is used as a mechanism for process adherence in complex tasks and does not attempt to bypass system safety filters.
[PROMPT_INJECTION]: The skill presents an indirect prompt injection surface by design, as it must ingest and process contents from the codebase.
Ingestion points: The skill reads existing files and search output using grep and glob (SKILL.md).
Boundary markers: Absent; the instructions do not specify the use of delimiters or provide warnings to ignore potential instructions embedded within the codebase files.
Capability inventory: The agent can execute local scripts, perform file system searches, and write analysis documentation to the .ai/workspace/analysis/ directory.
Sanitization: There is no evidence of input validation or content sanitization for the data read from the local files.

handoff