The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill uses the $ARGUMENTS variable to ingest user requests, which are then interpolated into the instructions provided to specialized sub-agents.
Ingestion points: The $ARGUMENTS variable is used to define the core query in SKILL.md and to launch the vibe-extras:codebase-explorer sub-agents.
Boundary markers: The skill does not employ explicit delimiters (e.g., XML tags or triple quotes) or 'ignore embedded instructions' warnings when passing user data to the sub-agent prompts.
Capability inventory: The skill possesses the ability to execute shell commands (date), write orchestration logs to the /tmp directory, and orchestrate external agents.
Sanitization: While the skill instructs the agent to format a 'topic-slug', there is no explicit requirement to sanitize or escape the user input before it is used in sub-agent prompts.
Remediation: Implement strict boundary markers around user-provided content and include explicit instructions for sub-agents to treat the interpolated text as data rather than instructions.
[COMMAND_EXECUTION]: The skill relies on shell command execution to generate metadata for its internal orchestration process.
Evidence: The instruction Run: date +%Y%m%d-%H%M%S in Phase 1.1 of SKILL.md triggers a local command execution to create a timestamp for file naming.

explore-codebase