The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes arbitrary commands through the repro-rerun script to verify bug reproductions. This capability is central to its purpose but allows the execution of shell commands based on inputs derived during the debugging session.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data such as bug reports, stack traces, and repository wiki hits to diagnose issues. These inputs are used to rank hypotheses and construct reproduction commands, creating a surface where malicious content in a bug report could influence agent behavior.\n
Ingestion points: User-provided <input> (bug reports, traces) and handoff_context.wiki_hits.\n
Boundary markers: The instructions do not define specific delimiters or warnings to ignore instructions within the processed data are defined.\n
Capability inventory: Executes shell commands via repro-rerun, performs file system reads/writes, and interacts with external tools like gh and /briesearch.\n
Sanitization: There are no instructions for sanitizing or escaping the content of bug reports before using them to generate reproduction commands or hypotheses.\n- [DYNAMIC_EXECUTION]: The skill utilizes a bundled Python Zip Application (pasteurize.pyz) to perform operational tasks such as rerunning reproduction commands and scanning for debugging tags in the codebase.

pasteurize