The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill executes remote setup scripts from the author's GitHub repository (github.com/Rheinmir/setup) using patterns like curl | bash and git clone followed by shell execution. These scripts are used for environment bootstrapping and installing internal tools from the vendor's infrastructure.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted codebase files in Phase 1 (Step 1.3) by passing content to the opencode tool for LLM-based batch analysis. This ingestion point lacks explicit boundary markers or sanitization, creating an attack surface where a malicious codebase could attempt to inject instructions into the agent's context during the analysis phase.
[COMMAND_EXECUTION]: The skill relies on extensive local command execution, including bash and python, to scan directories, extract git history, and render documentation. It also utilizes npx to manage dependencies and run a local web server for document previewing.
[DATA_EXFILTRATION]: The 'Output Report' phase includes a mandatory routine that clones the author's repository and uses git push to upload the local SKILL.md file. While documented as a synchronization mechanism, this involves network activity that exfiltrates local file content and state to a remote repository.

orca-onboard