The Agent Skills Directory

[SAFE]: The skill functions as a purely educational and methodological framework for finding and fixing software bugs. It does not contain any executable scripts, command-line instructions that could lead to unauthorized access, or remote connections to untrusted sources. The methodology emphasizes evidence-based investigation, systematic verification of hypotheses, and post-mortem analysis, all of which are standard engineering practices.
[INDIRECT_PROMPT_INJECTION]: While the workflow involves the analysis of external data such as logs and bug reports, the skill defines a logical reasoning process for the agent rather than providing vulnerable automated prompts. Ingestion points include error messages and logs described in the reproduction and investigation phases. Capability inventory is limited to standard tools (Read, Grep, Glob, Bash) used for diagnostic purposes. Boundary markers and sanitization are not explicitly defined as the skill focuses on instructional methodology rather than code implementation.

debugging