The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it analyzes untrusted content from the repository to make installation and classification decisions.
Ingestion points: Reads files using rg and cat, including package.json, pyproject.toml, and general source code files to establish the project stack.
Boundary markers: The instructions lack explicit boundary markers or directions to ignore instructions embedded within the data it reads, although it mandates citing concrete evidence for decisions.
Capability inventory: The skill has the capability to execute shell commands (rg, cat) and perform file-system modifications, specifically installing components into the .claude/skills/ directory.
Sanitization: No sanitization or validation of the repository content is specified before the agent processes it to generate the install plan.

agent-sort