The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes various system commands using git, gh (GitHub CLI), pnpm, and uv to automate branch creation, committing, pushing, and pull request generation.
[EXTERNAL_DOWNLOADS]: The skill performs automated package management and dependency synchronization using pnpm install, uv sync, and pnpm add. These operations fetch software resources from well-known public registries including npm and PyPI.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection via the processing of external tool outputs used to trigger auto-fixes.
Ingestion points: The skill monitors the standard output and error streams of lint, type-check, and test processes to detect specific error patterns.
Boundary markers: There are no identified markers or delimiters used to separate tool-generated output from the agent's instructional context.
Capability inventory: The skill possesses the capability to modify the local environment through package installations (pnpm add), code alterations (ruff --fix), and repository updates (git push).
Sanitization: No validation or sanitization is performed on the error strings before they are interpolated into auto-fix commands, potentially allowing malformed output to influence subsequent command execution.

blueprint-engine