The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The documentation includes an installation command npx skills add flyingsquirrel0419/squirrel-skill which fetches the skill definition from a remote GitHub repository.
[COMMAND_EXECUTION]: The skill's 'Test' and 'Polish' phases are designed to execute local development commands such as test runners, linters, and formatters to verify and improve code quality.
[PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it ingests and processes untrusted data from project files (source code, READMEs, and tests) which could contain adversarial instructions.
Ingestion points: Reads files in the local project directory during the 'Discover', 'Build', and 'Bug Hunt' phases.
Boundary markers: None explicitly defined in the instruction set to separate processed code from agent instructions.
Capability inventory: Performs file system reads/writes and shell command execution (tests/linting).
Sanitization: No evidence of input sanitization for the ingested code content before processing.

squirrel