The Agent Skills Directory

[SAFE]: The skill is a workflow guide that emphasizes development best practices such as requiring issue approval, establishing clean reproductions, and maintaining strict line-change limits (400 lines). It explicitly disclaims authority for repository management actions like merging or labeling, which limits its potential for misuse.- [INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources such as issues and pull requests, creating a potential attack surface for indirect instructions. 1. Ingestion points: The skill interacts with GitHub issues and pull requests to audit existing PRs and verify defect descriptions. 2. Boundary markers: No specific delimiters or instructions to ignore embedded prompts are provided in the workflow. 3. Capability inventory: The skill calls for running tests, mapping code with CodeGraph, and performing source-mutating normalization. 4. Sanitization: The workflow mitigates injection risks by requiring a verifiable reproduction on the main branch before implementation, ensuring that inputs are validated through functional behavior.

rdd-defect-workflow